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
  31 /**
  32  * Internationalisation code
  33  * @ingroup Language
  34  */
  35 class Language {
  36         /**
  37          * Return autonyms in fetchLanguageName(s).
  38          * @since 1.32
  39          */
  40         const AS_AUTONYMS = null;
  41
  42         /**
  43          * Return all known languages in fetchLanguageName(s).
  44          * @since 1.32
  45          */
  46         const ALL = 'all';
  47
  48         /**
  49          * Return in fetchLanguageName(s) only the languages for which we have at
  50          * least some localisation.
  51          * @since 1.32
  52          */
  53         const SUPPORTED = 'mwfile';
  54
  55         /**
  56          * @var LanguageConverter
  57          */
  58         public $mConverter;
  59
  60         public $mVariants, $mCode, $mLoaded = false;
  61         public $mMagicExtensions = [], $mMagicHookDone = false;
  62         private $mHtmlCode = null, $mParentLanguage = false;
  63
  64         public $dateFormatStrings = [];
  65         public $mExtendedSpecialPageAliases;
  66
  67         /** @var array|null */
  68         protected $namespaceNames;
  69         protected $mNamespaceIds, $namespaceAliases;
  70
  71         /**
  72          * ReplacementArray object caches
  73          */
  74         public $transformData = [];
  75
  76         /**
  77          * @var LocalisationCache
  78          */
  79         static public $dataCache;
  80
  81         static public $mLangObjCache = [];
  82
  83         static public $mWeekdayMsgs = [
  84                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  85                 'friday', 'saturday'
  86         ];
  87
  88         static public $mWeekdayAbbrevMsgs = [
  89                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
  90         ];
  91
  92         static public $mMonthMsgs = [
  93                 'january', 'february', 'march', 'april', 'may_long', 'june',
  94                 'july', 'august', 'september', 'october', 'november',
  95                 'december'
  96         ];
  97         static public $mMonthGenMsgs = [
  98                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
  99                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 100                 'december-gen'
 101         ];
 102         static public $mMonthAbbrevMsgs = [
 103                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 104                 'sep', 'oct', 'nov', 'dec'
 105         ];
 106
 107         static public $mIranianCalendarMonthMsgs = [
 108                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 109                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 110                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 111                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 112         ];
 113
 114         static public $mHebrewCalendarMonthMsgs = [
 115                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 116                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 117                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 118                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 119                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 120         ];
 121
 122         static public $mHebrewCalendarMonthGenMsgs = [
 123                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 124                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 125                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 126                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 127                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 128         ];
 129
 130         static public $mHijriCalendarMonthMsgs = [
 131                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 132                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 133                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 134                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 135         ];
 136
 137         /**
 138          * @since 1.20
 139          * @var array
 140          */
 141         static public $durationIntervals = [
 142                 'millennia' => 31556952000,
 143                 'centuries' => 3155695200,
 144                 'decades' => 315569520,
 145                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 146                 'weeks' => 604800,
 147                 'days' => 86400,
 148                 'hours' => 3600,
 149                 'minutes' => 60,
 150                 'seconds' => 1,
 151         ];
 152
 153         /**
 154          * Cache for language fallbacks.
 155          * @see Language::getFallbacksIncludingSiteLanguage
 156          * @since 1.21
 157          * @var array
 158          */
 159         static private $fallbackLanguageCache = [];
 160
 161         /**
 162          * Cache for grammar rules data
 163          * @var MapCacheLRU|null
 164          */
 165         static private $grammarTransformations;
 166
 167         /**
 168          * Cache for language names
 169          * @var HashBagOStuff|null
 170          */
 171         static private $languageNameCache;
 172
 173         /**
 174          * Unicode directional formatting characters, for embedBidi()
 175          */
 176         static private $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 177         static private $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 178         static private $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 179
 180         /**
 181          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 182          * - in group 1 if it is LTR
 183          * - in group 2 if it is RTL
 184          * Does not match if there is no strong directionality codepoint.
 185          *
 186          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 187          *
 188          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 189          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 190          */
 191         // @codeCoverageIgnoreStart
 192         // phpcs:ignore Generic.Files.LineLength
 193         static private $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';
 194         // @codeCoverageIgnoreEnd
 195
 196         /**
 197          * Get a cached or new language object for a given language code
 198          * @param string $code
 199          * @throws MWException
 200          * @return Language
 201          */
 202         static function factory( $code ) {
 203                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 204
 205                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 206                         $code = $wgDummyLanguageCodes[$code];
 207                 }
 208
 209                 // get the language object to process
 210                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 211
 212                 // merge the language object in to get it up front in the cache
 213                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 214                 // get rid of the oldest ones in case we have an overflow
 215                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 216
 217                 return $langObj;
 218         }
 219
 220         /**
 221          * Create a language object for a given language code
 222          * @param string $code
 223          * @param bool $fallback Whether we're going through language fallback chain
 224          * @throws MWException
 225          * @return Language
 226          */
 227         protected static function newFromCode( $code, $fallback = false ) {
 228                 if ( !self::isValidCode( $code ) ) {
 229                         throw new MWException( "Invalid language code \"$code\"" );
 230                 }
 231
 232                 if ( !self::isValidBuiltInCode( $code ) ) {
 233                         // It's not possible to customise this code with class files, so
 234                         // just return a Language object. This is to support uselang= hacks.
 235                         $lang = new Language;
 236                         $lang->setCode( $code );
 237                         return $lang;
 238                 }
 239
 240                 // Check if there is a language class for the code
 241                 $class = self::classFromCode( $code, $fallback );
 242                 // LanguageCode does not inherit Language
 243                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 244                         $lang = new $class;
 245                         return $lang;
 246                 }
 247
 248                 // Keep trying the fallback list until we find an existing class
 249                 $fallbacks = self::getFallbacksFor( $code );
 250                 foreach ( $fallbacks as $fallbackCode ) {
 251                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 252                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 253                         }
 254
 255                         $class = self::classFromCode( $fallbackCode );
 256                         if ( class_exists( $class ) ) {
 257                                 $lang = new $class;
 258                                 $lang->setCode( $code );
 259                                 return $lang;
 260                         }
 261                 }
 262
 263                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 264         }
 265
 266         /**
 267          * Intended for tests that may change configuration in a way that invalidates caches.
 268          *
 269          * @since 1.32
 270          */
 271         public static function clearCaches() {
 272                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 273                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 274                 }
 275                 self::$dataCache = null;
 276                 // Reinitialize $dataCache, since it's expected to always be available
 277                 self::getLocalisationCache();
 278                 self::$mLangObjCache = [];
 279                 self::$fallbackLanguageCache = [];
 280                 self::$grammarTransformations = null;
 281                 self::$languageNameCache = null;
 282         }
 283
 284         /**
 285          * Checks whether any localisation is available for that language tag
 286          * in MediaWiki (MessagesXx.php exists).
 287          *
 288          * @param string $code Language tag (in lower case)
 289          * @return bool Whether language is supported
 290          * @since 1.21
 291          */
 292         public static function isSupportedLanguage( $code ) {
 293                 if ( !self::isValidBuiltInCode( $code ) ) {
 294                         return false;
 295                 }
 296
 297                 if ( $code === 'qqq' ) {
 298                         return false;
 299                 }
 300
 301                 return is_readable( self::getMessagesFileName( $code ) ) ||
 302                         is_readable( self::getJsonMessagesFileName( $code ) );
 303         }
 304
 305         /**
 306          * Returns true if a language code string is a well-formed language tag
 307          * according to RFC 5646.
 308          * This function only checks well-formedness; it doesn't check that
 309          * language, script or variant codes actually exist in the repositories.
 310          *
 311          * Based on regexes by Mark Davis of the Unicode Consortium:
 312          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 313          *
 314          * @param string $code
 315          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 316          *
 317          * @return bool
 318          * @since 1.21
 319          */
 320         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 321                 $alpha = '[a-z]';
 322                 $digit = '[0-9]';
 323                 $alphanum = '[a-z0-9]';
 324                 $x = 'x'; # private use singleton
 325                 $singleton = '[a-wy-z]'; # other singleton
 326                 $s = $lenient ? '[-_]' : '-';
 327
 328                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 329                 $script = "$alpha{4}"; # ISO 15924
 330                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 331                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 332                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 333                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 334
 335                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 336                 # Since these are limited, this is safe even later changes to the registry --
 337                 # the only oddity is that it might change the type of the tag, and thus
 338                 # the results from the capturing groups.
 339                 # https://www.iana.org/assignments/language-subtag-registry
 340
 341                 $grandfathered = "en{$s}GB{$s}oed"
 342                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 343                         . "|no{$s}(?:bok|nyn)"
 344                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 345                         . "|zh{$s}min{$s}nan";
 346
 347                 $variantList = "$variant(?:$s$variant)*";
 348                 $extensionList = "$extension(?:$s$extension)*";
 349
 350                 $langtag = "(?:($language)"
 351                         . "(?:$s$script)?"
 352                         . "(?:$s$region)?"
 353                         . "(?:$s$variantList)?"
 354                         . "(?:$s$extensionList)?"
 355                         . "(?:$s$privateUse)?)";
 356
 357                 # The final breakdown, with capturing groups for each of these components
 358                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 359
 360                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 361
 362                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 363         }
 364
 365         /**
 366          * Returns true if a language code string is of a valid form, whether or
 367          * not it exists. This includes codes which are used solely for
 368          * customisation via the MediaWiki namespace.
 369          *
 370          * @param string $code
 371          *
 372          * @return bool
 373          */
 374         public static function isValidCode( $code ) {
 375                 static $cache = [];
 376                 if ( !isset( $cache[$code] ) ) {
 377                         // People think language codes are html safe, so enforce it.
 378                         // Ideally we should only allow a-zA-Z0-9-
 379                         // but, .+ and other chars are often used for {{int:}} hacks
 380                         // see bugs T39564, T39587, T38938
 381                         $cache[$code] =
 382                                 // Protect against path traversal
 383                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 384                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 385                 }
 386                 return $cache[$code];
 387         }
 388
 389         /**
 390          * Returns true if a language code is of a valid form for the purposes of
 391          * internal customisation of MediaWiki, via Messages*.php or *.json.
 392          *
 393          * @param string $code
 394          *
 395          * @throws MWException
 396          * @since 1.18
 397          * @return bool
 398          */
 399         public static function isValidBuiltInCode( $code ) {
 400                 if ( !is_string( $code ) ) {
 401                         if ( is_object( $code ) ) {
 402                                 $addmsg = " of class " . get_class( $code );
 403                         } else {
 404                                 $addmsg = '';
 405                         }
 406                         $type = gettype( $code );
 407                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 408                 }
 409
 410                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 411         }
 412
 413         /**
 414          * Returns true if a language code is an IETF tag known to MediaWiki.
 415          *
 416          * @param string $tag
 417          *
 418          * @since 1.21
 419          * @return bool
 420          */
 421         public static function isKnownLanguageTag( $tag ) {
 422                 // Quick escape for invalid input to avoid exceptions down the line
 423                 // when code tries to process tags which are not valid at all.
 424                 if ( !self::isValidBuiltInCode( $tag ) ) {
 425                         return false;
 426                 }
 427
 428                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 429                         || self::fetchLanguageName( $tag, $tag ) !== ''
 430                 ) {
 431                         return true;
 432                 }
 433
 434                 return false;
 435         }
 436
 437         /**
 438          * Get the LocalisationCache instance
 439          *
 440          * @return LocalisationCache
 441          */
 442         public static function getLocalisationCache() {
 443                 if ( is_null( self::$dataCache ) ) {
 444                         global $wgLocalisationCacheConf;
 445                         $class = $wgLocalisationCacheConf['class'];
 446                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 447                 }
 448                 return self::$dataCache;
 449         }
 450
 451         function __construct() {
 452                 $this->mConverter = new FakeConverter( $this );
 453                 // Set the code to the name of the descendant
 454                 if ( static::class === 'Language' ) {
 455                         $this->mCode = 'en';
 456                 } else {
 457                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 458                 }
 459                 self::getLocalisationCache();
 460         }
 461
 462         /**
 463          * Reduce memory usage
 464          */
 465         function __destruct() {
 466                 foreach ( $this as $name => $value ) {
 467                         unset( $this->$name );
 468                 }
 469         }
 470
 471         /**
 472          * Hook which will be called if this is the content language.
 473          * Descendants can use this to register hook functions or modify globals
 474          */
 475         function initContLang() {
 476         }
 477
 478         /**
 479          * @return array
 480          * @since 1.19
 481          */
 482         public function getFallbackLanguages() {
 483                 return self::getFallbacksFor( $this->mCode );
 484         }
 485
 486         /**
 487          * Exports $wgBookstoreListEn
 488          * @return array
 489          */
 490         public function getBookstoreList() {
 491                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 492         }
 493
 494         /**
 495          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 496          * available in localised form, it will be included in English.
 497          *
 498          * @return array
 499          */
 500         public function getNamespaces() {
 501                 if ( is_null( $this->namespaceNames ) ) {
 502                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 503
 504                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 505
 506                         $this->namespaceNames = $wgExtraNamespaces +
 507                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 508                         $this->namespaceNames += $validNamespaces;
 509
 510                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 511                         if ( $wgMetaNamespaceTalk ) {
 512                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 513                         } else {
 514                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 515                                 $this->namespaceNames[NS_PROJECT_TALK] =
 516                                         $this->fixVariableInNamespace( $talk );
 517                         }
 518
 519                         # Sometimes a language will be localised but not actually exist on this wiki.
 520                         foreach ( $this->namespaceNames as $key => $text ) {
 521                                 if ( !isset( $validNamespaces[$key] ) ) {
 522                                         unset( $this->namespaceNames[$key] );
 523                                 }
 524                         }
 525
 526                         # The above mixing may leave namespaces out of canonical order.
 527                         # Re-order by namespace ID number...
 528                         ksort( $this->namespaceNames );
 529
 530                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 531                 }
 532
 533                 return $this->namespaceNames;
 534         }
 535
 536         /**
 537          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 538          * @param array $namespaces Array of namespaces (id => name)
 539          */
 540         public function setNamespaces( array $namespaces ) {
 541                 $this->namespaceNames = $namespaces;
 542                 $this->mNamespaceIds = null;
 543         }
 544
 545         /**
 546          * Resets all of the namespace caches. Mainly used for testing
 547          */
 548         public function resetNamespaces() {
 549                 $this->namespaceNames = null;
 550                 $this->mNamespaceIds = null;
 551                 $this->namespaceAliases = null;
 552         }
 553
 554         /**
 555          * A convenience function that returns getNamespaces() with spaces instead of underscores
 556          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 557          *
 558          * @return array
 559          */
 560         public function getFormattedNamespaces() {
 561                 $ns = $this->getNamespaces();
 562                 foreach ( $ns as $k => $v ) {
 563                         $ns[$k] = strtr( $v, '_', ' ' );
 564                 }
 565                 return $ns;
 566         }
 567
 568         /**
 569          * Get a namespace value by key
 570          *
 571          * <code>
 572          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 573          * echo $mw_ns; // prints 'MediaWiki'
 574          * </code>
 575          *
 576          * @param int $index The array key of the namespace to return
 577          * @return string|bool String if the namespace value exists, otherwise false
 578          */
 579         public function getNsText( $index ) {
 580                 $ns = $this->getNamespaces();
 581                 return $ns[$index] ?? false;
 582         }
 583
 584         /**
 585          * A convenience function that returns the same thing as
 586          * getNsText() except with '_' changed to ' ', useful for
 587          * producing output.
 588          *
 589          * <code>
 590          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 591          * echo $mw_ns; // prints 'MediaWiki talk'
 592          * </code>
 593          *
 594          * @param int $index The array key of the namespace to return
 595          * @return string Namespace name without underscores (empty string if namespace does not exist)
 596          */
 597         public function getFormattedNsText( $index ) {
 598                 $ns = $this->getNsText( $index );
 599                 return strtr( $ns, '_', ' ' );
 600         }
 601
 602         /**
 603          * Returns gender-dependent namespace alias if available.
 604          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 605          * @param int $index Namespace index
 606          * @param string $gender Gender key (male, female... )
 607          * @return string
 608          * @since 1.18
 609          */
 610         public function getGenderNsText( $index, $gender ) {
 611                 global $wgExtraGenderNamespaces;
 612
 613                 $ns = $wgExtraGenderNamespaces +
 614                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 615
 616                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 617         }
 618
 619         /**
 620          * Whether this language uses gender-dependent namespace aliases.
 621          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 622          * @return bool
 623          * @since 1.18
 624          */
 625         public function needsGenderDistinction() {
 626                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 627                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 628                         // $wgExtraGenderNamespaces overrides everything
 629                         return true;
 630                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 631                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 632                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 633                         return false;
 634                 } else {
 635                         // Check what is in i18n files
 636                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 637                         return count( $aliases ) > 0;
 638                 }
 639         }
 640
 641         /**
 642          * Get a namespace key by value, case insensitive.
 643          * Only matches namespace names for the current language, not the
 644          * canonical ones defined in Namespace.php.
 645          *
 646          * @param string $text
 647          * @return int|bool An integer if $text is a valid value otherwise false
 648          */
 649         function getLocalNsIndex( $text ) {
 650                 $lctext = $this->lc( $text );
 651                 $ids = $this->getNamespaceIds();
 652                 return $ids[$lctext] ?? false;
 653         }
 654
 655         /**
 656          * @return array
 657          */
 658         public function getNamespaceAliases() {
 659                 if ( is_null( $this->namespaceAliases ) ) {
 660                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 661                         if ( !$aliases ) {
 662                                 $aliases = [];
 663                         } else {
 664                                 foreach ( $aliases as $name => $index ) {
 665                                         if ( $index === NS_PROJECT_TALK ) {
 666                                                 unset( $aliases[$name] );
 667                                                 $name = $this->fixVariableInNamespace( $name );
 668                                                 $aliases[$name] = $index;
 669                                         }
 670                                 }
 671                         }
 672
 673                         global $wgExtraGenderNamespaces;
 674                         $genders = $wgExtraGenderNamespaces +
 675                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 676                         foreach ( $genders as $index => $forms ) {
 677                                 foreach ( $forms as $alias ) {
 678                                         $aliases[$alias] = $index;
 679                                 }
 680                         }
 681
 682                         # Also add converted namespace names as aliases, to avoid confusion.
 683                         $convertedNames = [];
 684                         foreach ( $this->getVariants() as $variant ) {
 685                                 if ( $variant === $this->mCode ) {
 686                                         continue;
 687                                 }
 688                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 689                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 690                                 }
 691                         }
 692
 693                         $this->namespaceAliases = $aliases + $convertedNames;
 694                 }
 695
 696                 return $this->namespaceAliases;
 697         }
 698
 699         /**
 700          * @return array
 701          */
 702         public function getNamespaceIds() {
 703                 if ( is_null( $this->mNamespaceIds ) ) {
 704                         global $wgNamespaceAliases;
 705                         # Put namespace names and aliases into a hashtable.
 706                         # If this is too slow, then we should arrange it so that it is done
 707                         # before caching. The catch is that at pre-cache time, the above
 708                         # class-specific fixup hasn't been done.
 709                         $this->mNamespaceIds = [];
 710                         foreach ( $this->getNamespaces() as $index => $name ) {
 711                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 712                         }
 713                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 714                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 715                         }
 716                         if ( $wgNamespaceAliases ) {
 717                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 718                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 719                                 }
 720                         }
 721                 }
 722                 return $this->mNamespaceIds;
 723         }
 724
 725         /**
 726          * Get a namespace key by value, case insensitive.  Canonical namespace
 727          * names override custom ones defined for the current language.
 728          *
 729          * @param string $text
 730          * @return int|bool An integer if $text is a valid value otherwise false
 731          */
 732         public function getNsIndex( $text ) {
 733                 $lctext = $this->lc( $text );
 734                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 735                 if ( $ns !== null ) {
 736                         return $ns;
 737                 }
 738                 $ids = $this->getNamespaceIds();
 739                 return $ids[$lctext] ?? false;
 740         }
 741
 742         /**
 743          * short names for language variants used for language conversion links.
 744          *
 745          * @param string $code
 746          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 747          * @return string
 748          */
 749         public function getVariantname( $code, $usemsg = true ) {
 750                 $msg = "variantname-$code";
 751                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 752                         return $this->getMessageFromDB( $msg );
 753                 }
 754                 $name = self::fetchLanguageName( $code );
 755                 if ( $name ) {
 756                         return $name; # if it's defined as a language name, show that
 757                 } else {
 758                         # otherwise, output the language code
 759                         return $code;
 760                 }
 761         }
 762
 763         /**
 764          * @return string[]|bool List of date format preference keys, or false if disabled.
 765          */
 766         public function getDatePreferences() {
 767                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 768         }
 769
 770         /**
 771          * @return array
 772          */
 773         function getDateFormats() {
 774                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 775         }
 776
 777         /**
 778          * @return array|string
 779          */
 780         public function getDefaultDateFormat() {
 781                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 782                 if ( $df === 'dmy or mdy' ) {
 783                         global $wgAmericanDates;
 784                         return $wgAmericanDates ? 'mdy' : 'dmy';
 785                 } else {
 786                         return $df;
 787                 }
 788         }
 789
 790         /**
 791          * @return array
 792          */
 793         public function getDatePreferenceMigrationMap() {
 794                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 795         }
 796
 797         /**
 798          * @param string $image
 799          * @return array|null
 800          */
 801         function getImageFile( $image ) {
 802                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 803         }
 804
 805         /**
 806          * @return array
 807          * @since 1.24
 808          */
 809         public function getImageFiles() {
 810                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 811         }
 812
 813         /**
 814          * @return array
 815          */
 816         public function getExtraUserToggles() {
 817                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 818         }
 819
 820         /**
 821          * @param string $tog
 822          * @return string
 823          */
 824         function getUserToggle( $tog ) {
 825                 return $this->getMessageFromDB( "tog-$tog" );
 826         }
 827
 828         /**
 829          * Get an array of language names, indexed by code.
 830          * @param null|string $inLanguage Code of language in which to return the names
 831          *              Use self::AS_AUTONYMS for autonyms (native names)
 832          * @param string $include One of:
 833          *              self::ALL all available languages
 834          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 835          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 836          * @return array Language code => language name (sorted by key)
 837          * @since 1.20
 838          */
 839         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 840                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 841                 $cacheKey .= ":$include";
 842                 if ( self::$languageNameCache === null ) {
 843                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 844                 }
 845
 846                 $ret = self::$languageNameCache->get( $cacheKey );
 847                 if ( !$ret ) {
 848                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 849                         self::$languageNameCache->set( $cacheKey, $ret );
 850                 }
 851                 return $ret;
 852         }
 853
 854         /**
 855          * Uncached helper for fetchLanguageNames
 856          * @param null|string $inLanguage Code of language in which to return the names
 857          *              Use self::AS_AUTONYMS for autonyms (native names)
 858          * @param string $include One of:
 859          *              self::ALL all available languages
 860          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 861          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 862          * @return array Language code => language name (sorted by key)
 863          */
 864         private static function fetchLanguageNamesUncached(
 865                 $inLanguage = self::AS_AUTONYMS,
 866                 $include = 'mw'
 867         ) {
 868                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 869
 870                 // If passed an invalid language code to use, fallback to en
 871                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 872                         $inLanguage = 'en';
 873                 }
 874
 875                 $names = [];
 876
 877                 if ( $inLanguage ) {
 878                         # TODO: also include when $inLanguage is null, when this code is more efficient
 879                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 880                 }
 881
 882                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 883                 if ( $wgUsePigLatinVariant ) {
 884                         // Pig Latin (for variant development)
 885                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 886                 }
 887
 888                 foreach ( $mwNames as $mwCode => $mwName ) {
 889                         # - Prefer own MediaWiki native name when not using the hook
 890                         # - For other names just add if not added through the hook
 891                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 892                                 $names[$mwCode] = $mwName;
 893                         }
 894                 }
 895
 896                 if ( $include === self::ALL ) {
 897                         ksort( $names );
 898                         return $names;
 899                 }
 900
 901                 $returnMw = [];
 902                 $coreCodes = array_keys( $mwNames );
 903                 foreach ( $coreCodes as $coreCode ) {
 904                         $returnMw[$coreCode] = $names[$coreCode];
 905                 }
 906
 907                 if ( $include === self::SUPPORTED ) {
 908                         $namesMwFile = [];
 909                         # We do this using a foreach over the codes instead of a directory
 910                         # loop so that messages files in extensions will work correctly.
 911                         foreach ( $returnMw as $code => $value ) {
 912                                 if ( is_readable( self::getMessagesFileName( $code ) )
 913                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 914                                 ) {
 915                                         $namesMwFile[$code] = $names[$code];
 916                                 }
 917                         }
 918
 919                         ksort( $namesMwFile );
 920                         return $namesMwFile;
 921                 }
 922
 923                 ksort( $returnMw );
 924                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 925                 return $returnMw;
 926         }
 927
 928         /**
 929          * @param string $code The code of the language for which to get the name
 930          * @param null|string $inLanguage Code of language in which to return the name
 931          *   (SELF::AS_AUTONYMS for autonyms)
 932          * @param string $include See fetchLanguageNames()
 933          * @return string Language name or empty
 934          * @since 1.20
 935          */
 936         public static function fetchLanguageName(
 937                 $code,
 938                 $inLanguage = self::AS_AUTONYMS,
 939                 $include = self::ALL
 940         ) {
 941                 $code = strtolower( $code );
 942                 $array = self::fetchLanguageNames( $inLanguage, $include );
 943                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 944         }
 945
 946         /**
 947          * Get a message from the MediaWiki namespace.
 948          *
 949          * @param string $msg Message name
 950          * @return string
 951          */
 952         public function getMessageFromDB( $msg ) {
 953                 return $this->msg( $msg )->text();
 954         }
 955
 956         /**
 957          * Get message object in this language. Only for use inside this class.
 958          *
 959          * @param string $msg Message name
 960          * @return Message
 961          */
 962         protected function msg( $msg ) {
 963                 return wfMessage( $msg )->inLanguage( $this );
 964         }
 965
 966         /**
 967          * @param string $key
 968          * @return string
 969          */
 970         public function getMonthName( $key ) {
 971                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 972         }
 973
 974         /**
 975          * @return array
 976          */
 977         public function getMonthNamesArray() {
 978                 $monthNames = [ '' ];
 979                 for ( $i = 1; $i < 13; $i++ ) {
 980                         $monthNames[] = $this->getMonthName( $i );
 981                 }
 982                 return $monthNames;
 983         }
 984
 985         /**
 986          * @param string $key
 987          * @return string
 988          */
 989         public function getMonthNameGen( $key ) {
 990                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 991         }
 992
 993         /**
 994          * @param string $key
 995          * @return string
 996          */
 997         public function getMonthAbbreviation( $key ) {
 998                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 999         }
1000
1001         /**
1002          * @return array
1003          */
1004         public function getMonthAbbreviationsArray() {
1005                 $monthNames = [ '' ];
1006                 for ( $i = 1; $i < 13; $i++ ) {
1007                         $monthNames[] = $this->getMonthAbbreviation( $i );
1008                 }
1009                 return $monthNames;
1010         }
1011
1012         /**
1013          * @param string $key
1014          * @return string
1015          */
1016         public function getWeekdayName( $key ) {
1017                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1018         }
1019
1020         /**
1021          * @param string $key
1022          * @return string
1023          */
1024         function getWeekdayAbbreviation( $key ) {
1025                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1026         }
1027
1028         /**
1029          * @param string $key
1030          * @return string
1031          */
1032         function getIranianCalendarMonthName( $key ) {
1033                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1034         }
1035
1036         /**
1037          * @param string $key
1038          * @return string
1039          */
1040         function getHebrewCalendarMonthName( $key ) {
1041                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1042         }
1043
1044         /**
1045          * @param string $key
1046          * @return string
1047          */
1048         function getHebrewCalendarMonthNameGen( $key ) {
1049                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1050         }
1051
1052         /**
1053          * @param string $key
1054          * @return string
1055          */
1056         function getHijriCalendarMonthName( $key ) {
1057                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1058         }
1059
1060         /**
1061          * Pass through result from $dateTimeObj->format()
1062          * @param DateTime|bool|null &$dateTimeObj
1063          * @param string $ts
1064          * @param DateTimeZone|bool|null $zone
1065          * @param string $code
1066          * @return string
1067          */
1068         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1069                 if ( !$dateTimeObj ) {
1070                         $dateTimeObj = DateTime::createFromFormat(
1071                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1072                         );
1073                 }
1074                 return $dateTimeObj->format( $code );
1075         }
1076
1077         /**
1078          * This is a workalike of PHP's date() function, but with better
1079          * internationalisation, a reduced set of format characters, and a better
1080          * escaping format.
1081          *
1082          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1083          * the PHP manual for definitions. There are a number of extensions, which
1084          * start with "x":
1085          *
1086          *    xn   Do not translate digits of the next numeric format character
1087          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1088          *    xr   Use roman numerals for the next numeric format character
1089          *    xh   Use hebrew numerals for the next numeric format character
1090          *    xx   Literal x
1091          *    xg   Genitive month name
1092          *
1093          *    xij  j (day number) in Iranian calendar
1094          *    xiF  F (month name) in Iranian calendar
1095          *    xin  n (month number) in Iranian calendar
1096          *    xiy  y (two digit year) in Iranian calendar
1097          *    xiY  Y (full year) in Iranian calendar
1098          *    xit  t (days in month) in Iranian calendar
1099          *    xiz  z (day of the year) in Iranian calendar
1100          *
1101          *    xjj  j (day number) in Hebrew calendar
1102          *    xjF  F (month name) in Hebrew calendar
1103          *    xjt  t (days in month) in Hebrew calendar
1104          *    xjx  xg (genitive month name) in Hebrew calendar
1105          *    xjn  n (month number) in Hebrew calendar
1106          *    xjY  Y (full year) in Hebrew calendar
1107          *
1108          *    xmj  j (day number) in Hijri calendar
1109          *    xmF  F (month name) in Hijri calendar
1110          *    xmn  n (month number) in Hijri calendar
1111          *    xmY  Y (full year) in Hijri calendar
1112          *
1113          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1114          *                       identical to the Gregorian calendar
1115          *    xoY  Y (full year) in Minguo calendar or Juche year.
1116          *                       Months and days are identical to the
1117          *                       Gregorian calendar
1118          *    xtY  Y (full year) in Japanese nengo. Months and days are
1119          *                       identical to the Gregorian calendar
1120          *
1121          * Characters enclosed in double quotes will be considered literal (with
1122          * the quotes themselves removed). Unmatched quotes will be considered
1123          * literal quotes. Example:
1124          *
1125          * "The month is" F       => The month is January
1126          * i's"                   => 20'11"
1127          *
1128          * Backslash escaping is also supported.
1129          *
1130          * Input timestamp is assumed to be pre-normalized to the desired local
1131          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1132          * $ts is UTC if $zone is not given.
1133          *
1134          * @param string $format
1135          * @param string $ts 14-character timestamp
1136          *      YYYYMMDDHHMMSS
1137          *      01234567890123
1138          * @param DateTimeZone|null $zone Timezone of $ts
1139          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1140          * Only makes sense if $ts is the current time.
1141          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1142          *
1143          * @throws MWException
1144          * @return string
1145          */
1146         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1147                 $s = '';
1148                 $raw = false;
1149                 $roman = false;
1150                 $hebrewNum = false;
1151                 $dateTimeObj = false;
1152                 $rawToggle = false;
1153                 $iranian = false;
1154                 $hebrew = false;
1155                 $hijri = false;
1156                 $thai = false;
1157                 $minguo = false;
1158                 $tenno = false;
1159
1160                 $usedSecond = false;
1161                 $usedMinute = false;
1162                 $usedHour = false;
1163                 $usedAMPM = false;
1164                 $usedDay = false;
1165                 $usedWeek = false;
1166                 $usedMonth = false;
1167                 $usedYear = false;
1168                 $usedISOYear = false;
1169                 $usedIsLeapYear = false;
1170
1171                 $usedHebrewMonth = false;
1172                 $usedIranianMonth = false;
1173                 $usedHijriMonth = false;
1174                 $usedHebrewYear = false;
1175                 $usedIranianYear = false;
1176                 $usedHijriYear = false;
1177                 $usedTennoYear = false;
1178
1179                 if ( strlen( $ts ) !== 14 ) {
1180                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1181                 }
1182
1183                 if ( !ctype_digit( $ts ) ) {
1184                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1185                 }
1186
1187                 $formatLength = strlen( $format );
1188                 for ( $p = 0; $p < $formatLength; $p++ ) {
1189                         $num = false;
1190                         $code = $format[$p];
1191                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1192                                 $code .= $format[++$p];
1193                         }
1194
1195                         if ( ( $code === 'xi'
1196                                         || $code === 'xj'
1197                                         || $code === 'xk'
1198                                         || $code === 'xm'
1199                                         || $code === 'xo'
1200                                         || $code === 'xt' )
1201                                 && $p < $formatLength - 1 ) {
1202                                 $code .= $format[++$p];
1203                         }
1204
1205                         switch ( $code ) {
1206                                 case 'xx':
1207                                         $s .= 'x';
1208                                         break;
1209                                 case 'xn':
1210                                         $raw = true;
1211                                         break;
1212                                 case 'xN':
1213                                         $rawToggle = !$rawToggle;
1214                                         break;
1215                                 case 'xr':
1216                                         $roman = true;
1217                                         break;
1218                                 case 'xh':
1219                                         $hebrewNum = true;
1220                                         break;
1221                                 case 'xg':
1222                                         $usedMonth = true;
1223                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1224                                         break;
1225                                 case 'xjx':
1226                                         $usedHebrewMonth = true;
1227                                         if ( !$hebrew ) {
1228                                                 $hebrew = self::tsToHebrew( $ts );
1229                                         }
1230                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1231                                         break;
1232                                 case 'd':
1233                                         $usedDay = true;
1234                                         $num = substr( $ts, 6, 2 );
1235                                         break;
1236                                 case 'D':
1237                                         $usedDay = true;
1238                                         $s .= $this->getWeekdayAbbreviation(
1239                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1240                                         );
1241                                         break;
1242                                 case 'j':
1243                                         $usedDay = true;
1244                                         $num = intval( substr( $ts, 6, 2 ) );
1245                                         break;
1246                                 case 'xij':
1247                                         $usedDay = true;
1248                                         if ( !$iranian ) {
1249                                                 $iranian = self::tsToIranian( $ts );
1250                                         }
1251                                         $num = $iranian[2];
1252                                         break;
1253                                 case 'xmj':
1254                                         $usedDay = true;
1255                                         if ( !$hijri ) {
1256                                                 $hijri = self::tsToHijri( $ts );
1257                                         }
1258                                         $num = $hijri[2];
1259                                         break;
1260                                 case 'xjj':
1261                                         $usedDay = true;
1262                                         if ( !$hebrew ) {
1263                                                 $hebrew = self::tsToHebrew( $ts );
1264                                         }
1265                                         $num = $hebrew[2];
1266                                         break;
1267                                 case 'l':
1268                                         $usedDay = true;
1269                                         $s .= $this->getWeekdayName(
1270                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1271                                         );
1272                                         break;
1273                                 case 'F':
1274                                         $usedMonth = true;
1275                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1276                                         break;
1277                                 case 'xiF':
1278                                         $usedIranianMonth = true;
1279                                         if ( !$iranian ) {
1280                                                 $iranian = self::tsToIranian( $ts );
1281                                         }
1282                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1283                                         break;
1284                                 case 'xmF':
1285                                         $usedHijriMonth = true;
1286                                         if ( !$hijri ) {
1287                                                 $hijri = self::tsToHijri( $ts );
1288                                         }
1289                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1290                                         break;
1291                                 case 'xjF':
1292                                         $usedHebrewMonth = true;
1293                                         if ( !$hebrew ) {
1294                                                 $hebrew = self::tsToHebrew( $ts );
1295                                         }
1296                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1297                                         break;
1298                                 case 'm':
1299                                         $usedMonth = true;
1300                                         $num = substr( $ts, 4, 2 );
1301                                         break;
1302                                 case 'M':
1303                                         $usedMonth = true;
1304                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1305                                         break;
1306                                 case 'n':
1307                                         $usedMonth = true;
1308                                         $num = intval( substr( $ts, 4, 2 ) );
1309                                         break;
1310                                 case 'xin':
1311                                         $usedIranianMonth = true;
1312                                         if ( !$iranian ) {
1313                                                 $iranian = self::tsToIranian( $ts );
1314                                         }
1315                                         $num = $iranian[1];
1316                                         break;
1317                                 case 'xmn':
1318                                         $usedHijriMonth = true;
1319                                         if ( !$hijri ) {
1320                                                 $hijri = self::tsToHijri( $ts );
1321                                         }
1322                                         $num = $hijri[1];
1323                                         break;
1324                                 case 'xjn':
1325                                         $usedHebrewMonth = true;
1326                                         if ( !$hebrew ) {
1327                                                 $hebrew = self::tsToHebrew( $ts );
1328                                         }
1329                                         $num = $hebrew[1];
1330                                         break;
1331                                 case 'xjt':
1332                                         $usedHebrewMonth = true;
1333                                         if ( !$hebrew ) {
1334                                                 $hebrew = self::tsToHebrew( $ts );
1335                                         }
1336                                         $num = $hebrew[3];
1337                                         break;
1338                                 case 'Y':
1339                                         $usedYear = true;
1340                                         $num = substr( $ts, 0, 4 );
1341                                         break;
1342                                 case 'xiY':
1343                                         $usedIranianYear = true;
1344                                         if ( !$iranian ) {
1345                                                 $iranian = self::tsToIranian( $ts );
1346                                         }
1347                                         $num = $iranian[0];
1348                                         break;
1349                                 case 'xmY':
1350                                         $usedHijriYear = true;
1351                                         if ( !$hijri ) {
1352                                                 $hijri = self::tsToHijri( $ts );
1353                                         }
1354                                         $num = $hijri[0];
1355                                         break;
1356                                 case 'xjY':
1357                                         $usedHebrewYear = true;
1358                                         if ( !$hebrew ) {
1359                                                 $hebrew = self::tsToHebrew( $ts );
1360                                         }
1361                                         $num = $hebrew[0];
1362                                         break;
1363                                 case 'xkY':
1364                                         $usedYear = true;
1365                                         if ( !$thai ) {
1366                                                 $thai = self::tsToYear( $ts, 'thai' );
1367                                         }
1368                                         $num = $thai[0];
1369                                         break;
1370                                 case 'xoY':
1371                                         $usedYear = true;
1372                                         if ( !$minguo ) {
1373                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1374                                         }
1375                                         $num = $minguo[0];
1376                                         break;
1377                                 case 'xtY':
1378                                         $usedTennoYear = true;
1379                                         if ( !$tenno ) {
1380                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1381                                         }
1382                                         $num = $tenno[0];
1383                                         break;
1384                                 case 'y':
1385                                         $usedYear = true;
1386                                         $num = substr( $ts, 2, 2 );
1387                                         break;
1388                                 case 'xiy':
1389                                         $usedIranianYear = true;
1390                                         if ( !$iranian ) {
1391                                                 $iranian = self::tsToIranian( $ts );
1392                                         }
1393                                         $num = substr( $iranian[0], -2 );
1394                                         break;
1395                                 case 'xit':
1396                                         $usedIranianYear = true;
1397                                         if ( !$iranian ) {
1398                                                 $iranian = self::tsToIranian( $ts );
1399                                         }
1400                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1401                                         break;
1402                                 case 'xiz':
1403                                         $usedIranianYear = true;
1404                                         if ( !$iranian ) {
1405                                                 $iranian = self::tsToIranian( $ts );
1406                                         }
1407                                         $num = $iranian[3];
1408                                         break;
1409                                 case 'a':
1410                                         $usedAMPM = true;
1411                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1412                                         break;
1413                                 case 'A':
1414                                         $usedAMPM = true;
1415                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1416                                         break;
1417                                 case 'g':
1418                                         $usedHour = true;
1419                                         $h = substr( $ts, 8, 2 );
1420                                         $num = $h % 12 ? $h % 12 : 12;
1421                                         break;
1422                                 case 'G':
1423                                         $usedHour = true;
1424                                         $num = intval( substr( $ts, 8, 2 ) );
1425                                         break;
1426                                 case 'h':
1427                                         $usedHour = true;
1428                                         $h = substr( $ts, 8, 2 );
1429                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1430                                         break;
1431                                 case 'H':
1432                                         $usedHour = true;
1433                                         $num = substr( $ts, 8, 2 );
1434                                         break;
1435                                 case 'i':
1436                                         $usedMinute = true;
1437                                         $num = substr( $ts, 10, 2 );
1438                                         break;
1439                                 case 's':
1440                                         $usedSecond = true;
1441                                         $num = substr( $ts, 12, 2 );
1442                                         break;
1443                                 case 'c':
1444                                 case 'r':
1445                                         $usedSecond = true;
1446                                         // fall through
1447                                 case 'e':
1448                                 case 'O':
1449                                 case 'P':
1450                                 case 'T':
1451                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1452                                         break;
1453                                 case 'w':
1454                                 case 'N':
1455                                 case 'z':
1456                                         $usedDay = true;
1457                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1458                                         break;
1459                                 case 'W':
1460                                         $usedWeek = true;
1461                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1462                                         break;
1463                                 case 't':
1464                                         $usedMonth = true;
1465                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1466                                         break;
1467                                 case 'L':
1468                                         $usedIsLeapYear = true;
1469                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1470                                         break;
1471                                 case 'o':
1472                                         $usedISOYear = true;
1473                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1474                                         break;
1475                                 case 'U':
1476                                         $usedSecond = true;
1477                                         // fall through
1478                                 case 'I':
1479                                 case 'Z':
1480                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1481                                         break;
1482                                 case '\\':
1483                                         # Backslash escaping
1484                                         if ( $p < $formatLength - 1 ) {
1485                                                 $s .= $format[++$p];
1486                                         } else {
1487                                                 $s .= '\\';
1488                                         }
1489                                         break;
1490                                 case '"':
1491                                         # Quoted literal
1492                                         if ( $p < $formatLength - 1 ) {
1493                                                 $endQuote = strpos( $format, '"', $p + 1 );
1494                                                 if ( $endQuote === false ) {
1495                                                         # No terminating quote, assume literal "
1496                                                         $s .= '"';
1497                                                 } else {
1498                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1499                                                         $p = $endQuote;
1500                                                 }
1501                                         } else {
1502                                                 # Quote at end of string, assume literal "
1503                                                 $s .= '"';
1504                                         }
1505                                         break;
1506                                 default:
1507                                         $s .= $format[$p];
1508                         }
1509                         if ( $num !== false ) {
1510                                 if ( $rawToggle || $raw ) {
1511                                         $s .= $num;
1512                                         $raw = false;
1513                                 } elseif ( $roman ) {
1514                                         $s .= self::romanNumeral( $num );
1515                                         $roman = false;
1516                                 } elseif ( $hebrewNum ) {
1517                                         $s .= self::hebrewNumeral( $num );
1518                                         $hebrewNum = false;
1519                                 } else {
1520                                         $s .= $this->formatNum( $num, true );
1521                                 }
1522                         }
1523                 }
1524
1525                 if ( $ttl === 'unused' ) {
1526                         // No need to calculate the TTL, the caller wont use it anyway.
1527                 } elseif ( $usedSecond ) {
1528                         $ttl = 1;
1529                 } elseif ( $usedMinute ) {
1530                         $ttl = 60 - substr( $ts, 12, 2 );
1531                 } elseif ( $usedHour ) {
1532                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1533                 } elseif ( $usedAMPM ) {
1534                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1535                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1536                 } elseif (
1537                         $usedDay ||
1538                         $usedHebrewMonth ||
1539                         $usedIranianMonth ||
1540                         $usedHijriMonth ||
1541                         $usedHebrewYear ||
1542                         $usedIranianYear ||
1543                         $usedHijriYear ||
1544                         $usedTennoYear
1545                 ) {
1546                         // @todo Someone who understands the non-Gregorian calendars
1547                         // should write proper logic for them so that they don't need purged every day.
1548                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1549                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1550                 } else {
1551                         $possibleTtls = [];
1552                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1553                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1554                         if ( $usedWeek ) {
1555                                 $possibleTtls[] =
1556                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1557                                         $timeRemainingInDay;
1558                         } elseif ( $usedISOYear ) {
1559                                 // December 28th falls on the last ISO week of the year, every year.
1560                                 // The last ISO week of a year can be 52 or 53.
1561                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1562                                         'Ymd',
1563                                         substr( $ts, 0, 4 ) . '1228',
1564                                         $zone ?: new DateTimeZone( 'UTC' )
1565                                 )->format( 'W' );
1566                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1567                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1568                                 $timeRemainingInWeek =
1569                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1570                                         + $timeRemainingInDay;
1571                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1572                         }
1573
1574                         if ( $usedMonth ) {
1575                                 $possibleTtls[] =
1576                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1577                                                 substr( $ts, 6, 2 ) ) * 86400
1578                                         + $timeRemainingInDay;
1579                         } elseif ( $usedYear ) {
1580                                 $possibleTtls[] =
1581                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1582                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1583                                         + $timeRemainingInDay;
1584                         } elseif ( $usedIsLeapYear ) {
1585                                 $year = substr( $ts, 0, 4 );
1586                                 $timeRemainingInYear =
1587                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1588                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1589                                         + $timeRemainingInDay;
1590                                 $mod = $year % 4;
1591                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1592                                         // this isn't a leap year. see when the next one starts
1593                                         $nextCandidate = $year - $mod + 4;
1594                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1595                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1596                                                         $timeRemainingInYear;
1597                                         } else {
1598                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1599                                                         $timeRemainingInYear;
1600                                         }
1601                                 } else {
1602                                         // this is a leap year, so the next year isn't
1603                                         $possibleTtls[] = $timeRemainingInYear;
1604                                 }
1605                         }
1606
1607                         if ( $possibleTtls ) {
1608                                 $ttl = min( $possibleTtls );
1609                         }
1610                 }
1611
1612                 return $s;
1613         }
1614
1615         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1616         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1617
1618         /**
1619          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1620          * Gregorian dates to Iranian dates. Originally written in C, it
1621          * is released under the terms of GNU Lesser General Public
1622          * License. Conversion to PHP was performed by Niklas Laxström.
1623          *
1624          * Link: http://www.farsiweb.info/jalali/jalali.c
1625          *
1626          * @param string $ts
1627          *
1628          * @return int[]
1629          */
1630         private static function tsToIranian( $ts ) {
1631                 $gy = substr( $ts, 0, 4 ) - 1600;
1632                 $gm = substr( $ts, 4, 2 ) - 1;
1633                 $gd = substr( $ts, 6, 2 ) - 1;
1634
1635                 # Days passed from the beginning (including leap years)
1636                 $gDayNo = 365 * $gy
1637                         + floor( ( $gy + 3 ) / 4 )
1638                         - floor( ( $gy + 99 ) / 100 )
1639                         + floor( ( $gy + 399 ) / 400 );
1640
1641                 // Add days of the past months of this year
1642                 for ( $i = 0; $i < $gm; $i++ ) {
1643                         $gDayNo += self::$GREG_DAYS[$i];
1644                 }
1645
1646                 // Leap years
1647                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1648                         $gDayNo++;
1649                 }
1650
1651                 // Days passed in current month
1652                 $gDayNo += (int)$gd;
1653
1654                 $jDayNo = $gDayNo - 79;
1655
1656                 $jNp = floor( $jDayNo / 12053 );
1657                 $jDayNo %= 12053;
1658
1659                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1660                 $jDayNo %= 1461;
1661
1662                 if ( $jDayNo >= 366 ) {
1663                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1664                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1665                 }
1666
1667                 $jz = $jDayNo;
1668
1669                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1670                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1671                 }
1672
1673                 $jm = $i + 1;
1674                 $jd = $jDayNo + 1;
1675
1676                 return [ $jy, $jm, $jd, $jz ];
1677         }
1678
1679         /**
1680          * Converting Gregorian dates to Hijri dates.
1681          *
1682          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1683          *
1684          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1685          *
1686          * @param string $ts
1687          *
1688          * @return int[]
1689          */
1690         private static function tsToHijri( $ts ) {
1691                 $year = substr( $ts, 0, 4 );
1692                 $month = substr( $ts, 4, 2 );
1693                 $day = substr( $ts, 6, 2 );
1694
1695                 $zyr = $year;
1696                 $zd = $day;
1697                 $zm = $month;
1698                 $zy = $zyr;
1699
1700                 if (
1701                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1702                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1703                 ) {
1704                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1705                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1706                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1707                                         $zd - 32075;
1708                 } else {
1709                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1710                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1711                 }
1712
1713                 $zl = $zjd - 1948440 + 10632;
1714                 $zn = (int)( ( $zl - 1 ) / 10631 );
1715                 $zl = $zl - 10631 * $zn + 354;
1716                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1717                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1718                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1719                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1720                 $zm = (int)( ( 24 * $zl ) / 709 );
1721                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1722                 $zy = 30 * $zn + $zj - 30;
1723
1724                 return [ $zy, $zm, $zd ];
1725         }
1726
1727         /**
1728          * Converting Gregorian dates to Hebrew dates.
1729          *
1730          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1731          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1732          * to translate the relevant functions into PHP and release them under
1733          * GNU GPL.
1734          *
1735          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1736          * and Adar II is 14. In a non-leap year, Adar is 6.
1737          *
1738          * @param string $ts
1739          *
1740          * @return int[]
1741          */
1742         private static function tsToHebrew( $ts ) {
1743                 # Parse date
1744                 $year = substr( $ts, 0, 4 );
1745                 $month = substr( $ts, 4, 2 );
1746                 $day = substr( $ts, 6, 2 );
1747
1748                 # Calculate Hebrew year
1749                 $hebrewYear = $year + 3760;
1750
1751                 # Month number when September = 1, August = 12
1752                 $month += 4;
1753                 if ( $month > 12 ) {
1754                         # Next year
1755                         $month -= 12;
1756                         $year++;
1757                         $hebrewYear++;
1758                 }
1759
1760                 # Calculate day of year from 1 September
1761                 $dayOfYear = $day;
1762                 for ( $i = 1; $i < $month; $i++ ) {
1763                         if ( $i == 6 ) {
1764                                 # February
1765                                 $dayOfYear += 28;
1766                                 # Check if the year is leap
1767                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1768                                         $dayOfYear++;
1769                                 }
1770                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1771                                 $dayOfYear += 30;
1772                         } else {
1773                                 $dayOfYear += 31;
1774                         }
1775                 }
1776
1777                 # Calculate the start of the Hebrew year
1778                 $start = self::hebrewYearStart( $hebrewYear );
1779
1780                 # Calculate next year's start
1781                 if ( $dayOfYear <= $start ) {
1782                         # Day is before the start of the year - it is the previous year
1783                         # Next year's start
1784                         $nextStart = $start;
1785                         # Previous year
1786                         $year--;
1787                         $hebrewYear--;
1788                         # Add days since previous year's 1 September
1789                         $dayOfYear += 365;
1790                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1791                                 # Leap year
1792                                 $dayOfYear++;
1793                         }
1794                         # Start of the new (previous) year
1795                         $start = self::hebrewYearStart( $hebrewYear );
1796                 } else {
1797                         # Next year's start
1798                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1799                 }
1800
1801                 # Calculate Hebrew day of year
1802                 $hebrewDayOfYear = $dayOfYear - $start;
1803
1804                 # Difference between year's days
1805                 $diff = $nextStart - $start;
1806                 # Add 12 (or 13 for leap years) days to ignore the difference between
1807                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1808                 # difference is only about the year type
1809                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1810                         $diff += 13;
1811                 } else {
1812                         $diff += 12;
1813                 }
1814
1815                 # Check the year pattern, and is leap year
1816                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1817                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1818                 # and non-leap years
1819                 $yearPattern = $diff % 30;
1820                 # Check if leap year
1821                 $isLeap = $diff >= 30;
1822
1823                 # Calculate day in the month from number of day in the Hebrew year
1824                 # Don't check Adar - if the day is not in Adar, we will stop before;
1825                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1826                 $hebrewDay = $hebrewDayOfYear;
1827                 $hebrewMonth = 1;
1828                 $days = 0;
1829                 while ( $hebrewMonth <= 12 ) {
1830                         # Calculate days in this month
1831                         if ( $isLeap && $hebrewMonth == 6 ) {
1832                                 # Adar in a leap year
1833                                 if ( $isLeap ) {
1834                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1835                                         $days = 30;
1836                                         if ( $hebrewDay <= $days ) {
1837                                                 # Day in Adar I
1838                                                 $hebrewMonth = 13;
1839                                         } else {
1840                                                 # Subtract the days of Adar I
1841                                                 $hebrewDay -= $days;
1842                                                 # Try Adar II
1843                                                 $days = 29;
1844                                                 if ( $hebrewDay <= $days ) {
1845                                                         # Day in Adar II
1846                                                         $hebrewMonth = 14;
1847                                                 }
1848                                         }
1849                                 }
1850                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1851                                 # Cheshvan in a complete year (otherwise as the rule below)
1852                                 $days = 30;
1853                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1854                                 # Kislev in an incomplete year (otherwise as the rule below)
1855                                 $days = 29;
1856                         } else {
1857                                 # Odd months have 30 days, even have 29
1858                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1859                         }
1860                         if ( $hebrewDay <= $days ) {
1861                                 # In the current month
1862                                 break;
1863                         } else {
1864                                 # Subtract the days of the current month
1865                                 $hebrewDay -= $days;
1866                                 # Try in the next month
1867                                 $hebrewMonth++;
1868                         }
1869                 }
1870
1871                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1872         }
1873
1874         /**
1875          * This calculates the Hebrew year start, as days since 1 September.
1876          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1877          * Used for Hebrew date.
1878          *
1879          * @param int $year
1880          *
1881          * @return string
1882          */
1883         private static function hebrewYearStart( $year ) {
1884                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1885                 $b = intval( ( $year - 1 ) % 4 );
1886                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1887                 if ( $m < 0 ) {
1888                         $m--;
1889                 }
1890                 $Mar = intval( $m );
1891                 if ( $m < 0 ) {
1892                         $m++;
1893                 }
1894                 $m -= $Mar;
1895
1896                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1897                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1898                         $Mar++;
1899                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1900                         $Mar += 2;
1901                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1902                         $Mar++;
1903                 }
1904
1905                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1906                 return $Mar;
1907         }
1908
1909         /**
1910          * Algorithm to convert Gregorian dates to Thai solar dates,
1911          * Minguo dates or Minguo dates.
1912          *
1913          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1914          *       https://en.wikipedia.org/wiki/Minguo_calendar
1915          *       https://en.wikipedia.org/wiki/Japanese_era_name
1916          *
1917          * @param string $ts 14-character timestamp
1918          * @param string $cName Calender name
1919          * @return array Converted year, month, day
1920          */
1921         private static function tsToYear( $ts, $cName ) {
1922                 $gy = substr( $ts, 0, 4 );
1923                 $gm = substr( $ts, 4, 2 );
1924                 $gd = substr( $ts, 6, 2 );
1925
1926                 if ( !strcmp( $cName, 'thai' ) ) {
1927                         # Thai solar dates
1928                         # Add 543 years to the Gregorian calendar
1929                         # Months and days are identical
1930                         $gy_offset = $gy + 543;
1931                         # fix for dates between 1912 and 1941
1932                         # https://en.wikipedia.org/?oldid=836596673#New_year
1933                         if ( $gy >= 1912 && $gy <= 1940 ) {
1934                                 if ( $gm <= 3 ) {
1935                                         $gy_offset--;
1936                                 }
1937                                 $gm = ( $gm - 3 ) % 12;
1938                         }
1939                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1940                         # Minguo dates
1941                         # Deduct 1911 years from the Gregorian calendar
1942                         # Months and days are identical
1943                         $gy_offset = $gy - 1911;
1944                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1945                         # Nengō dates up to Meiji period
1946                         # Deduct years from the Gregorian calendar
1947                         # depending on the nengo periods
1948                         # Months and days are identical
1949                         if ( ( $gy < 1912 )
1950                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1951                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1952                         ) {
1953                                 # Meiji period
1954                                 $gy_gannen = $gy - 1868 + 1;
1955                                 $gy_offset = $gy_gannen;
1956                                 if ( $gy_gannen == 1 ) {
1957                                         $gy_offset = '元';
1958                                 }
1959                                 $gy_offset = '明治' . $gy_offset;
1960                         } elseif (
1961                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1962                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1963                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1964                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1965                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1966                         ) {
1967                                 # Taishō period
1968                                 $gy_gannen = $gy - 1912 + 1;
1969                                 $gy_offset = $gy_gannen;
1970                                 if ( $gy_gannen == 1 ) {
1971                                         $gy_offset = '元';
1972                                 }
1973                                 $gy_offset = '大正' . $gy_offset;
1974                         } elseif (
1975                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1976                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1977                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1978                         ) {
1979                                 # Shōwa period
1980                                 $gy_gannen = $gy - 1926 + 1;
1981                                 $gy_offset = $gy_gannen;
1982                                 if ( $gy_gannen == 1 ) {
1983                                         $gy_offset = '元';
1984                                 }
1985                                 $gy_offset = '昭和' . $gy_offset;
1986                         } else {
1987                                 # Heisei period
1988                                 $gy_gannen = $gy - 1989 + 1;
1989                                 $gy_offset = $gy_gannen;
1990                                 if ( $gy_gannen == 1 ) {
1991                                         $gy_offset = '元';
1992                                 }
1993                                 $gy_offset = '平成' . $gy_offset;
1994                         }
1995                 } else {
1996                         $gy_offset = $gy;
1997                 }
1998
1999                 return [ $gy_offset, $gm, $gd ];
2000         }
2001
2002         /**
2003          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2004          *
2005          * This is the rule the BIDI algorithm uses to determine the directionality of
2006          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2007          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2008          *
2009          * TODO: Does not handle BIDI control characters inside the text.
2010          * TODO: Does not handle unallocated characters.
2011          *
2012          * @param string $text Text to test
2013          * @return null|string Directionality ('ltr' or 'rtl') or null
2014          */
2015         private static function strongDirFromContent( $text = '' ) {
2016                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2017                         return null;
2018                 }
2019                 if ( $matches[1] === '' ) {
2020                         return 'rtl';
2021                 }
2022                 return 'ltr';
2023         }
2024
2025         /**
2026          * Roman number formatting up to 10000
2027          *
2028          * @param int $num
2029          *
2030          * @return string
2031          */
2032         static function romanNumeral( $num ) {
2033                 static $table = [
2034                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2035                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2036                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2037                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2038                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2039                 ];
2040
2041                 $num = intval( $num );
2042                 if ( $num > 10000 || $num <= 0 ) {
2043                         return $num;
2044                 }
2045
2046                 $s = '';
2047                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2048                         if ( $num >= $pow10 ) {
2049                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2050                         }
2051                         $num = $num % $pow10;
2052                 }
2053                 return $s;
2054         }
2055
2056         /**
2057          * Hebrew Gematria number formatting up to 9999
2058          *
2059          * @param int $num
2060          *
2061          * @return string
2062          */
2063         static function hebrewNumeral( $num ) {
2064                 static $table = [
2065                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2066                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2067                         [ '',
2068                                 [ 'ק' ],
2069                                 [ 'ר' ],
2070                                 [ 'ש' ],
2071                                 [ 'ת' ],
2072                                 [ 'ת', 'ק' ],
2073                                 [ 'ת', 'ר' ],
2074                                 [ 'ת', 'ש' ],
2075                                 [ 'ת', 'ת' ],
2076                                 [ 'ת', 'ת', 'ק' ],
2077                                 [ 'ת', 'ת', 'ר' ],
2078                         ],
2079                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2080                 ];
2081
2082                 $num = intval( $num );
2083                 if ( $num > 9999 || $num <= 0 ) {
2084                         return $num;
2085                 }
2086
2087                 // Round thousands have special notations
2088                 if ( $num === 1000 ) {
2089                         return "א' אלף";
2090                 } elseif ( $num % 1000 === 0 ) {
2091                         return $table[0][$num / 1000] . "' אלפים";
2092                 }
2093
2094                 $letters = [];
2095
2096                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2097                         if ( $num >= $pow10 ) {
2098                                 if ( $num === 15 || $num === 16 ) {
2099                                         $letters[] = $table[0][9];
2100                                         $letters[] = $table[0][$num - 9];
2101                                         $num = 0;
2102                                 } else {
2103                                         $letters = array_merge(
2104                                                 $letters,
2105                                                 (array)$table[$i][intval( $num / $pow10 )]
2106                                         );
2107
2108                                         if ( $pow10 === 1000 ) {
2109                                                 $letters[] = "'";
2110                                         }
2111                                 }
2112                         }
2113
2114                         $num = $num % $pow10;
2115                 }
2116
2117                 $preTransformLength = count( $letters );
2118                 if ( $preTransformLength === 1 ) {
2119                         // Add geresh (single quote) to one-letter numbers
2120                         $letters[] = "'";
2121                 } else {
2122                         $lastIndex = $preTransformLength - 1;
2123                         $letters[$lastIndex] = str_replace(
2124                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2125                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2126                                 $letters[$lastIndex]
2127                         );
2128
2129                         // Add gershayim (double quote) to multiple-letter numbers,
2130                         // but exclude numbers with only one letter after the thousands
2131                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2132                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2133                                 $letters[] = "'";
2134                         } else {
2135                                 array_splice( $letters, -1, 0, '"' );
2136                         }
2137                 }
2138
2139                 return implode( $letters );
2140         }
2141
2142         /**
2143          * Used by date() and time() to adjust the time output.
2144          *
2145          * @param string $ts The time in date('YmdHis') format
2146          * @param mixed $tz Adjust the time by this amount (default false, mean we
2147          *   get user timecorrection setting)
2148          * @return int
2149          */
2150         public function userAdjust( $ts, $tz = false ) {
2151                 global $wgUser, $wgLocalTZoffset;
2152
2153                 if ( $tz === false ) {
2154                         $tz = $wgUser->getOption( 'timecorrection' );
2155                 }
2156
2157                 $data = explode( '|', $tz, 3 );
2158
2159                 if ( $data[0] == 'ZoneInfo' ) {
2160                         try {
2161                                 $userTZ = new DateTimeZone( $data[2] );
2162                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2163                                 $date->setTimezone( $userTZ );
2164                                 return $date->format( 'YmdHis' );
2165                         } catch ( Exception $e ) {
2166                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2167                                 $data[0] = 'Offset';
2168                         }
2169                 }
2170
2171                 if ( $data[0] == 'System' || $tz == '' ) {
2172                         # Global offset in minutes.
2173                         $minDiff = $wgLocalTZoffset;
2174                 } elseif ( $data[0] == 'Offset' ) {
2175                         $minDiff = intval( $data[1] );
2176                 } else {
2177                         $data = explode( ':', $tz );
2178                         if ( count( $data ) == 2 ) {
2179                                 $data[0] = intval( $data[0] );
2180                                 $data[1] = intval( $data[1] );
2181                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2182                                 if ( $data[0] < 0 ) {
2183                                         $minDiff = -$minDiff;
2184                                 }
2185                         } else {
2186                                 $minDiff = intval( $data[0] ) * 60;
2187                         }
2188                 }
2189
2190                 # No difference ? Return time unchanged
2191                 if ( 0 == $minDiff ) {
2192                         return $ts;
2193                 }
2194
2195                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2196                 # Generate an adjusted date; take advantage of the fact that mktime
2197                 # will normalize out-of-range values so we don't have to split $minDiff
2198                 # into hours and minutes.
2199                 $t = mktime( (
2200                         (int)substr( $ts, 8, 2 ) ), # Hours
2201                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2202                         (int)substr( $ts, 12, 2 ), # Seconds
2203                         (int)substr( $ts, 4, 2 ), # Month
2204                         (int)substr( $ts, 6, 2 ), # Day
2205                         (int)substr( $ts, 0, 4 ) ); # Year
2206
2207                 $date = date( 'YmdHis', $t );
2208                 Wikimedia\restoreWarnings();
2209
2210                 return $date;
2211         }
2212
2213         /**
2214          * This is meant to be used by time(), date(), and timeanddate() to get
2215          * the date preference they're supposed to use, it should be used in
2216          * all children.
2217          *
2218          *     function timeanddate([...], $format = true) {
2219          *       $datePreference = $this->dateFormat($format);
2220          *       [...]
2221          *     }
2222          *
2223          * @param int|string|bool $usePrefs If true, the user's preference is used
2224          *   if false, the site/language default is used
2225          *   if int/string, assumed to be a format.
2226          * @return string
2227          */
2228         function dateFormat( $usePrefs = true ) {
2229                 global $wgUser;
2230
2231                 if ( is_bool( $usePrefs ) ) {
2232                         if ( $usePrefs ) {
2233                                 $datePreference = $wgUser->getDatePreference();
2234                         } else {
2235                                 $datePreference = (string)User::getDefaultOption( 'date' );
2236                         }
2237                 } else {
2238                         $datePreference = (string)$usePrefs;
2239                 }
2240
2241                 // return int
2242                 if ( $datePreference == '' ) {
2243                         return 'default';
2244                 }
2245
2246                 return $datePreference;
2247         }
2248
2249         /**
2250          * Get a format string for a given type and preference
2251          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2252          * @param string $pref The format name as it appears in Messages*.php under
2253          *  $datePreferences.
2254          *
2255          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2256          *
2257          * @return string
2258          */
2259         function getDateFormatString( $type, $pref ) {
2260                 $wasDefault = false;
2261                 if ( $pref == 'default' ) {
2262                         $wasDefault = true;
2263                         $pref = $this->getDefaultDateFormat();
2264                 }
2265
2266                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2267                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2268
2269                         if ( $type === 'pretty' && $df === null ) {
2270                                 $df = $this->getDateFormatString( 'date', $pref );
2271                         }
2272
2273                         if ( !$wasDefault && $df === null ) {
2274                                 $pref = $this->getDefaultDateFormat();
2275                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2276                         }
2277
2278                         $this->dateFormatStrings[$type][$pref] = $df;
2279                 }
2280                 return $this->dateFormatStrings[$type][$pref];
2281         }
2282
2283         /**
2284          * @param string $ts The time format which needs to be turned into a
2285          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2286          * @param bool $adj Whether to adjust the time output according to the
2287          *   user configured offset ($timecorrection)
2288          * @param mixed $format True to use user's date format preference
2289          * @param string|bool $timecorrection The time offset as returned by
2290          *   validateTimeZone() in Special:Preferences
2291          * @return string
2292          */
2293         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2294                 $ts = wfTimestamp( TS_MW, $ts );
2295                 if ( $adj ) {
2296                         $ts = $this->userAdjust( $ts, $timecorrection );
2297                 }
2298                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2299                 return $this->sprintfDate( $df, $ts );
2300         }
2301
2302         /**
2303          * @param string $ts The time format which needs to be turned into a
2304          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2305          * @param bool $adj Whether to adjust the time output according to the
2306          *   user configured offset ($timecorrection)
2307          * @param mixed $format True to use user's date format preference
2308          * @param string|bool $timecorrection The time offset as returned by
2309          *   validateTimeZone() in Special:Preferences
2310          * @return string
2311          */
2312         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2313                 $ts = wfTimestamp( TS_MW, $ts );
2314                 if ( $adj ) {
2315                         $ts = $this->userAdjust( $ts, $timecorrection );
2316                 }
2317                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2318                 return $this->sprintfDate( $df, $ts );
2319         }
2320
2321         /**
2322          * @param string $ts The time format which needs to be turned into a
2323          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2324          * @param bool $adj Whether to adjust the time output according to the
2325          *   user configured offset ($timecorrection)
2326          * @param mixed $format What format to return, if it's false output the
2327          *   default one (default true)
2328          * @param string|bool $timecorrection The time offset as returned by
2329          *   validateTimeZone() in Special:Preferences
2330          * @return string
2331          */
2332         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2333                 $ts = wfTimestamp( TS_MW, $ts );
2334                 if ( $adj ) {
2335                         $ts = $this->userAdjust( $ts, $timecorrection );
2336                 }
2337                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2338                 return $this->sprintfDate( $df, $ts );
2339         }
2340
2341         /**
2342          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2343          *
2344          * @since 1.20
2345          *
2346          * @param int $seconds The amount of seconds.
2347          * @param array $chosenIntervals The intervals to enable.
2348          *
2349          * @return string
2350          */
2351         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2352                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2353
2354                 $segments = [];
2355
2356                 foreach ( $intervals as $intervalName => $intervalValue ) {
2357                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2358                         // duration-years, duration-decades, duration-centuries, duration-millennia
2359                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2360                         $segments[] = $message->inLanguage( $this )->escaped();
2361                 }
2362
2363                 return $this->listToText( $segments );
2364         }
2365
2366         /**
2367          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2368          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2369          *
2370          * @since 1.20
2371          *
2372          * @param int $seconds The amount of seconds.
2373          * @param array $chosenIntervals The intervals to enable.
2374          *
2375          * @return array
2376          */
2377         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2378                 if ( empty( $chosenIntervals ) ) {
2379                         $chosenIntervals = [
2380                                 'millennia',
2381                                 'centuries',
2382                                 'decades',
2383                                 'years',
2384                                 'days',
2385                                 'hours',
2386                                 'minutes',
2387                                 'seconds'
2388                         ];
2389                 }
2390
2391                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2392                 $sortedNames = array_keys( $intervals );
2393                 $smallestInterval = array_pop( $sortedNames );
2394
2395                 $segments = [];
2396
2397                 foreach ( $intervals as $name => $length ) {
2398                         $value = floor( $seconds / $length );
2399
2400                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2401                                 $seconds -= $value * $length;
2402                                 $segments[$name] = $value;
2403                         }
2404                 }
2405
2406                 return $segments;
2407         }
2408
2409         /**
2410          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2411          *
2412          * @param string $type Can be 'date', 'time' or 'both'
2413          * @param string $ts The time format which needs to be turned into a
2414          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2415          * @param User $user User object used to get preferences for timezone and format
2416          * @param array $options Array, can contain the following keys:
2417          *   - 'timecorrection': time correction, can have the following values:
2418          *     - true: use user's preference
2419          *     - false: don't use time correction
2420          *     - int: value of time correction in minutes
2421          *   - 'format': format to use, can have the following values:
2422          *     - true: use user's preference
2423          *     - false: use default preference
2424          *     - string: format to use
2425          * @since 1.19
2426          * @return string
2427          */
2428         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2429                 $ts = wfTimestamp( TS_MW, $ts );
2430                 $options += [ 'timecorrection' => true, 'format' => true ];
2431                 if ( $options['timecorrection'] !== false ) {
2432                         if ( $options['timecorrection'] === true ) {
2433                                 $offset = $user->getOption( 'timecorrection' );
2434                         } else {
2435                                 $offset = $options['timecorrection'];
2436                         }
2437                         $ts = $this->userAdjust( $ts, $offset );
2438                 }
2439                 if ( $options['format'] === true ) {
2440                         $format = $user->getDatePreference();
2441                 } else {
2442                         $format = $options['format'];
2443                 }
2444                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2445                 return $this->sprintfDate( $df, $ts );
2446         }
2447
2448         /**
2449          * Get the formatted date for the given timestamp and formatted for
2450          * the given user.
2451          *
2452          * @param mixed $ts Mixed: the time format which needs to be turned into a
2453          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2454          * @param User $user User object used to get preferences for timezone and format
2455          * @param array $options Array, can contain the following keys:
2456          *   - 'timecorrection': time correction, can have the following values:
2457          *     - true: use user's preference
2458          *     - false: don't use time correction
2459          *     - int: value of time correction in minutes
2460          *   - 'format': format to use, can have the following values:
2461          *     - true: use user's preference
2462          *     - false: use default preference
2463          *     - string: format to use
2464          * @since 1.19
2465          * @return string
2466          */
2467         public function userDate( $ts, User $user, array $options = [] ) {
2468                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2469         }
2470
2471         /**
2472          * Get the formatted time for the given timestamp and formatted for
2473          * the given user.
2474          *
2475          * @param mixed $ts The time format which needs to be turned into a
2476          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2477          * @param User $user User object used to get preferences for timezone and format
2478          * @param array $options Array, can contain the following keys:
2479          *   - 'timecorrection': time correction, can have the following values:
2480          *     - true: use user's preference
2481          *     - false: don't use time correction
2482          *     - int: value of time correction in minutes
2483          *   - 'format': format to use, can have the following values:
2484          *     - true: use user's preference
2485          *     - false: use default preference
2486          *     - string: format to use
2487          * @since 1.19
2488          * @return string
2489          */
2490         public function userTime( $ts, User $user, array $options = [] ) {
2491                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2492         }
2493
2494         /**
2495          * Get the formatted date and time for the given timestamp and formatted for
2496          * the given user.
2497          *
2498          * @param mixed $ts The time format which needs to be turned into a
2499          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2500          * @param User $user User object used to get preferences for timezone and format
2501          * @param array $options Array, can contain the following keys:
2502          *   - 'timecorrection': time correction, can have the following values:
2503          *     - true: use user's preference
2504          *     - false: don't use time correction
2505          *     - int: value of time correction in minutes
2506          *   - 'format': format to use, can have the following values:
2507          *     - true: use user's preference
2508          *     - false: use default preference
2509          *     - string: format to use
2510          * @since 1.19
2511          * @return string
2512          */
2513         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2514                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2515         }
2516
2517         /**
2518          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2519          *
2520          * Determine the difference between the timestamp and the current time, and
2521          * generate a readable timestamp by returning "<N> <units> ago", where the
2522          * largest possible unit is used.
2523          *
2524          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2525          *
2526          * @param MWTimestamp $time
2527          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2528          * @param User|null $user User the timestamp is being generated for
2529          *  (or null to use main context's user)
2530          * @return string Formatted timestamp
2531          */
2532         public function getHumanTimestamp(
2533                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2534         ) {
2535                 if ( $relativeTo === null ) {
2536                         $relativeTo = new MWTimestamp();
2537                 }
2538                 if ( $user === null ) {
2539                         $user = RequestContext::getMain()->getUser();
2540                 }
2541
2542                 // Adjust for the user's timezone.
2543                 $offsetThis = $time->offsetForUser( $user );
2544                 $offsetRel = $relativeTo->offsetForUser( $user );
2545
2546                 $ts = '';
2547                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2548                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2549                 }
2550
2551                 // Reset the timezone on the objects.
2552                 $time->timestamp->sub( $offsetThis );
2553                 $relativeTo->timestamp->sub( $offsetRel );
2554
2555                 return $ts;
2556         }
2557
2558         /**
2559          * Convert an MWTimestamp into a pretty human-readable timestamp using
2560          * the given user preferences and relative base time.
2561          *
2562          * @see Language::getHumanTimestamp
2563          * @param MWTimestamp $ts Timestamp to prettify
2564          * @param MWTimestamp $relativeTo Base timestamp
2565          * @param User $user User preferences to use
2566          * @return string Human timestamp
2567          * @since 1.26
2568          */
2569         private function getHumanTimestampInternal(
2570                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2571         ) {
2572                 $diff = $ts->diff( $relativeTo );
2573                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2574                         (int)$relativeTo->timestamp->format( 'w' ) );
2575                 $days = $diff->days ?: (int)$diffDay;
2576                 if ( $diff->invert || $days > 5
2577                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2578                 ) {
2579                         // Timestamps are in different years: use full timestamp
2580                         // Also do full timestamp for future dates
2581                         /**
2582                          * @todo FIXME: Add better handling of future timestamps.
2583                          */
2584                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2585                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2586                 } elseif ( $days > 5 ) {
2587                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2588                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2589                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2590                 } elseif ( $days > 1 ) {
2591                         // Timestamp within the past week: show the day of the week and time
2592                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2593                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2594                         // Messages:
2595                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2596                         $ts = wfMessage( "$weekday-at" )
2597                                 ->inLanguage( $this )
2598                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2599                                 ->text();
2600                 } elseif ( $days == 1 ) {
2601                         // Timestamp was yesterday: say 'yesterday' and the time.
2602                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2603                         $ts = wfMessage( 'yesterday-at' )
2604                                 ->inLanguage( $this )
2605                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2606                                 ->text();
2607                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2608                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2609                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2610                         $ts = wfMessage( 'today-at' )
2611                                 ->inLanguage( $this )
2612                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2613                                 ->text();
2614
2615                 // From here on in, the timestamp was soon enough ago so that we can simply say
2616                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2617                 } elseif ( $diff->h == 1 ) {
2618                         // Less than 90 minutes, but more than an hour ago.
2619                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2620                 } elseif ( $diff->i >= 1 ) {
2621                         // A few minutes ago.
2622                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2623                 } elseif ( $diff->s >= 30 ) {
2624                         // Less than a minute, but more than 30 sec ago.
2625                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2626                 } else {
2627                         // Less than 30 seconds ago.
2628                         $ts = wfMessage( 'just-now' )->text();
2629                 }
2630
2631                 return $ts;
2632         }
2633
2634         /**
2635          * @param string $key
2636          * @return string|null
2637          */
2638         public function getMessage( $key ) {
2639                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2640         }
2641
2642         /**
2643          * @return array
2644          */
2645         function getAllMessages() {
2646                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2647         }
2648
2649         /**
2650          * @param string $in
2651          * @param string $out
2652          * @param string $string
2653          * @return string
2654          */
2655         public function iconv( $in, $out, $string ) {
2656                 # Even with //IGNORE iconv can whine about illegal characters in
2657                 # *input* string. We just ignore those too.
2658                 # REF: https://bugs.php.net/bug.php?id=37166
2659                 # REF: https://phabricator.wikimedia.org/T18885
2660                 Wikimedia\suppressWarnings();
2661                 $text = iconv( $in, $out . '//IGNORE', $string );
2662                 Wikimedia\restoreWarnings();
2663                 return $text;
2664         }
2665
2666         // callback functions for ucwords(), ucwordbreaks()
2667
2668         /**
2669          * @param array $matches
2670          * @return mixed|string
2671          */
2672         function ucwordbreaksCallbackAscii( $matches ) {
2673                 return $this->ucfirst( $matches[1] );
2674         }
2675
2676         /**
2677          * @param array $matches
2678          * @return string
2679          */
2680         function ucwordbreaksCallbackMB( $matches ) {
2681                 return mb_strtoupper( $matches[0] );
2682         }
2683
2684         /**
2685          * @param array $matches
2686          * @return string
2687          */
2688         function ucwordsCallbackMB( $matches ) {
2689                 return mb_strtoupper( $matches[0] );
2690         }
2691
2692         /**
2693          * Make a string's first character uppercase
2694          *
2695          * @param string $str
2696          *
2697          * @return string
2698          */
2699         public function ucfirst( $str ) {
2700                 $o = ord( $str );
2701                 if ( $o < 96 ) { // if already uppercase...
2702                         return $str;
2703                 } elseif ( $o < 128 ) {
2704                         return ucfirst( $str ); // use PHP's ucfirst()
2705                 } else {
2706                         // fall back to more complex logic in case of multibyte strings
2707                         return $this->uc( $str, true );
2708                 }
2709         }
2710
2711         /**
2712          * Convert a string to uppercase
2713          *
2714          * @param string $str
2715          * @param bool $first
2716          *
2717          * @return string
2718          */
2719         public function uc( $str, $first = false ) {
2720                 if ( $first ) {
2721                         if ( $this->isMultibyte( $str ) ) {
2722                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2723                         } else {
2724                                 return ucfirst( $str );
2725                         }
2726                 } else {
2727                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2728                 }
2729         }
2730
2731         /**
2732          * @param string $str
2733          * @return mixed|string
2734          */
2735         function lcfirst( $str ) {
2736                 $o = ord( $str );
2737                 if ( !$o ) {
2738                         return strval( $str );
2739                 } elseif ( $o >= 128 ) {
2740                         return $this->lc( $str, true );
2741                 } elseif ( $o > 96 ) {
2742                         return $str;
2743                 } else {
2744                         $str[0] = strtolower( $str[0] );
2745                         return $str;
2746                 }
2747         }
2748
2749         /**
2750          * @param string $str
2751          * @param bool $first
2752          * @return mixed|string
2753          */
2754         function lc( $str, $first = false ) {
2755                 if ( $first ) {
2756                         if ( $this->isMultibyte( $str ) ) {
2757                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2758                         } else {
2759                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2760                         }
2761                 } else {
2762                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2763                 }
2764         }
2765
2766         /**
2767          * @param string $str
2768          * @return bool
2769          */
2770         function isMultibyte( $str ) {
2771                 return strlen( $str ) !== mb_strlen( $str );
2772         }
2773
2774         /**
2775          * @param string $str
2776          * @return mixed|string
2777          */
2778         function ucwords( $str ) {
2779                 if ( $this->isMultibyte( $str ) ) {
2780                         $str = $this->lc( $str );
2781
2782                         // regexp to find first letter in each word (i.e. after each space)
2783                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2784
2785                         // function to use to capitalize a single char
2786                         return preg_replace_callback(
2787                                 $replaceRegexp,
2788                                 [ $this, 'ucwordsCallbackMB' ],
2789                                 $str
2790                         );
2791                 } else {
2792                         return ucwords( strtolower( $str ) );
2793                 }
2794         }
2795
2796         /**
2797          * capitalize words at word breaks
2798          *
2799          * @param string $str
2800          * @return mixed
2801          */
2802         function ucwordbreaks( $str ) {
2803                 if ( $this->isMultibyte( $str ) ) {
2804                         $str = $this->lc( $str );
2805
2806                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2807                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2808
2809                         // find first letter after word break
2810                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2811                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2812
2813                         return preg_replace_callback(
2814                                 $replaceRegexp,
2815                                 [ $this, 'ucwordbreaksCallbackMB' ],
2816                                 $str
2817                         );
2818                 } else {
2819                         return preg_replace_callback(
2820                                 '/\b([\w\x80-\xff]+)\b/',
2821                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2822                                 $str
2823                         );
2824                 }
2825         }
2826
2827         /**
2828          * Return a case-folded representation of $s
2829          *
2830          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2831          * and $s2 are the same except for the case of their characters. It is not
2832          * necessary for the value returned to make sense when displayed.
2833          *
2834          * Do *not* perform any other normalisation in this function. If a caller
2835          * uses this function when it should be using a more general normalisation
2836          * function, then fix the caller.
2837          *
2838          * @param string $s
2839          *
2840          * @return string
2841          */
2842         function caseFold( $s ) {
2843                 return $this->uc( $s );
2844         }
2845
2846         /**
2847          * @param string $s
2848          * @return string
2849          * @throws MWException
2850          */
2851         function checkTitleEncoding( $s ) {
2852                 if ( is_array( $s ) ) {
2853                         throw new MWException( 'Given array to checkTitleEncoding.' );
2854                 }
2855                 if ( StringUtils::isUtf8( $s ) ) {
2856                         return $s;
2857                 }
2858
2859                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2860         }
2861
2862         /**
2863          * @return array
2864          */
2865         function fallback8bitEncoding() {
2866                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2867         }
2868
2869         /**
2870          * Most writing systems use whitespace to break up words.
2871          * Some languages such as Chinese don't conventionally do this,
2872          * which requires special handling when breaking up words for
2873          * searching etc.
2874          *
2875          * @return bool
2876          */
2877         function hasWordBreaks() {
2878                 return true;
2879         }
2880
2881         /**
2882          * Some languages such as Chinese require word segmentation,
2883          * Specify such segmentation when overridden in derived class.
2884          *
2885          * @param string $string
2886          * @return string
2887          */
2888         function segmentByWord( $string ) {
2889                 return $string;
2890         }
2891
2892         /**
2893          * Some languages have special punctuation need to be normalized.
2894          * Make such changes here.
2895          *
2896          * @param string $string
2897          * @return string
2898          */
2899         function normalizeForSearch( $string ) {
2900                 return self::convertDoubleWidth( $string );
2901         }
2902
2903         /**
2904          * convert double-width roman characters to single-width.
2905          * range: ff00-ff5f ~= 0020-007f
2906          *
2907          * @param string $string
2908          *
2909          * @return string
2910          */
2911         protected static function convertDoubleWidth( $string ) {
2912                 static $full = null;
2913                 static $half = null;
2914
2915                 if ( $full === null ) {
2916                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2917                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2918                         $full = str_split( $fullWidth, 3 );
2919                         $half = str_split( $halfWidth );
2920                 }
2921
2922                 $string = str_replace( $full, $half, $string );
2923                 return $string;
2924         }
2925
2926         /**
2927          * @param string $string
2928          * @param string $pattern
2929          * @return string
2930          */
2931         protected static function insertSpace( $string, $pattern ) {
2932                 $string = preg_replace( $pattern, " $1 ", $string );
2933                 $string = preg_replace( '/ +/', ' ', $string );
2934                 return $string;
2935         }
2936
2937         /**
2938          * @param array $termsArray
2939          * @return array
2940          */
2941         function convertForSearchResult( $termsArray ) {
2942                 # some languages, e.g. Chinese, need to do a conversion
2943                 # in order for search results to be displayed correctly
2944                 return $termsArray;
2945         }
2946
2947         /**
2948          * Get the first character of a string.
2949          *
2950          * @param string $s
2951          * @return string
2952          */
2953         function firstChar( $s ) {
2954                 $matches = [];
2955                 preg_match(
2956                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2957                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2958                         $s,
2959                         $matches
2960                 );
2961
2962                 if ( isset( $matches[1] ) ) {
2963                         if ( strlen( $matches[1] ) != 3 ) {
2964                                 return $matches[1];
2965                         }
2966
2967                         // Break down Hangul syllables to grab the first jamo
2968                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2969                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2970                                 return $matches[1];
2971                         } elseif ( $code < 0xb098 ) {
2972                                 return "\u{3131}";
2973                         } elseif ( $code < 0xb2e4 ) {
2974                                 return "\u{3134}";
2975                         } elseif ( $code < 0xb77c ) {
2976                                 return "\u{3137}";
2977                         } elseif ( $code < 0xb9c8 ) {
2978                                 return "\u{3139}";
2979                         } elseif ( $code < 0xbc14 ) {
2980                                 return "\u{3141}";
2981                         } elseif ( $code < 0xc0ac ) {
2982                                 return "\u{3142}";
2983                         } elseif ( $code < 0xc544 ) {
2984                                 return "\u{3145}";
2985                         } elseif ( $code < 0xc790 ) {
2986                                 return "\u{3147}";
2987                         } elseif ( $code < 0xcc28 ) {
2988                                 return "\u{3148}";
2989                         } elseif ( $code < 0xce74 ) {
2990                                 return "\u{314A}";
2991                         } elseif ( $code < 0xd0c0 ) {
2992                                 return "\u{314B}";
2993                         } elseif ( $code < 0xd30c ) {
2994                                 return "\u{314C}";
2995                         } elseif ( $code < 0xd558 ) {
2996                                 return "\u{314D}";
2997                         } else {
2998                                 return "\u{314E}";
2999                         }
3000                 } else {
3001                         return '';
3002                 }
3003         }
3004
3005         /**
3006          * @deprecated No-op since 1.28
3007          */
3008         function initEncoding() {
3009                 wfDeprecated( __METHOD__, '1.28' );
3010                 // No-op.
3011         }
3012
3013         /**
3014          * @param string $s
3015          * @return string
3016          * @deprecated No-op since 1.28
3017          */
3018         function recodeForEdit( $s ) {
3019                 wfDeprecated( __METHOD__, '1.28' );
3020                 return $s;
3021         }
3022
3023         /**
3024          * @param string $s
3025          * @return string
3026          * @deprecated No-op since 1.28
3027          */
3028         function recodeInput( $s ) {
3029                 wfDeprecated( __METHOD__, '1.28' );
3030                 return $s;
3031         }
3032
3033         /**
3034          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3035          * also cleans up certain backwards-compatible sequences, converting them
3036          * to the modern Unicode equivalent.
3037          *
3038          * This is language-specific for performance reasons only.
3039          *
3040          * @param string $s
3041          *
3042          * @return string
3043          */
3044         public function normalize( $s ) {
3045                 global $wgAllUnicodeFixes;
3046                 $s = UtfNormal\Validator::cleanUp( $s );
3047                 if ( $wgAllUnicodeFixes ) {
3048                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3049                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3050                 }
3051
3052                 return $s;
3053         }
3054
3055         /**
3056          * Transform a string using serialized data stored in the given file (which
3057          * must be in the serialized subdirectory of $IP). The file contains pairs
3058          * mapping source characters to destination characters.
3059          *
3060          * The data is cached in process memory. This will go faster if you have the
3061          * FastStringSearch extension.
3062          *
3063          * @param string $file
3064          * @param string $string
3065          *
3066          * @throws MWException
3067          * @return string
3068          */
3069         protected function transformUsingPairFile( $file, $string ) {
3070                 if ( !isset( $this->transformData[$file] ) ) {
3071                         global $IP;
3072                         $data = require "$IP/languages/data/{$file}";
3073                         $this->transformData[$file] = new ReplacementArray( $data );
3074                 }
3075                 return $this->transformData[$file]->replace( $string );
3076         }
3077
3078         /**
3079          * For right-to-left language support
3080          *
3081          * @return bool
3082          */
3083         function isRTL() {
3084                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3085         }
3086
3087         /**
3088          * Return the correct HTML 'dir' attribute value for this language.
3089          * @return string
3090          */
3091         function getDir() {
3092                 return $this->isRTL() ? 'rtl' : 'ltr';
3093         }
3094
3095         /**
3096          * Return 'left' or 'right' as appropriate alignment for line-start
3097          * for this language's text direction.
3098          *
3099          * Should be equivalent to CSS3 'start' text-align value....
3100          *
3101          * @return string
3102          */
3103         function alignStart() {
3104                 return $this->isRTL() ? 'right' : 'left';
3105         }
3106
3107         /**
3108          * Return 'right' or 'left' as appropriate alignment for line-end
3109          * for this language's text direction.
3110          *
3111          * Should be equivalent to CSS3 'end' text-align value....
3112          *
3113          * @return string
3114          */
3115         function alignEnd() {
3116                 return $this->isRTL() ? 'left' : 'right';
3117         }
3118
3119         /**
3120          * A hidden direction mark (LRM or RLM), depending on the language direction.
3121          * Unlike getDirMark(), this function returns the character as an HTML entity.
3122          * This function should be used when the output is guaranteed to be HTML,
3123          * because it makes the output HTML source code more readable. When
3124          * the output is plain text or can be escaped, getDirMark() should be used.
3125          *
3126          * @param bool $opposite Get the direction mark opposite to your language
3127          * @return string
3128          * @since 1.20
3129          */
3130         function getDirMarkEntity( $opposite = false ) {
3131                 if ( $opposite ) {
3132                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3133                 }
3134                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3135         }
3136
3137         /**
3138          * A hidden direction mark (LRM or RLM), depending on the language direction.
3139          * This function produces them as invisible Unicode characters and
3140          * the output may be hard to read and debug, so it should only be used
3141          * when the output is plain text or can be escaped. When the output is
3142          * HTML, use getDirMarkEntity() instead.
3143          *
3144          * @param bool $opposite Get the direction mark opposite to your language
3145          * @return string
3146          */
3147         function getDirMark( $opposite = false ) {
3148                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3149                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3150                 if ( $opposite ) {
3151                         return $this->isRTL() ? $lrm : $rlm;
3152                 }
3153                 return $this->isRTL() ? $rlm : $lrm;
3154         }
3155
3156         /**
3157          * @return array
3158          */
3159         function capitalizeAllNouns() {
3160                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3161         }
3162
3163         /**
3164          * An arrow, depending on the language direction.
3165          *
3166          * @param string $direction The direction of the arrow: forwards (default),
3167          *   backwards, left, right, up, down.
3168          * @return string
3169          */
3170         function getArrow( $direction = 'forwards' ) {
3171                 switch ( $direction ) {
3172                         case 'forwards':
3173                                 return $this->isRTL() ? '←' : '→';
3174                         case 'backwards':
3175                                 return $this->isRTL() ? '→' : '←';
3176                         case 'left':
3177                                 return '←';
3178                         case 'right':
3179                                 return '→';
3180                         case 'up':
3181                                 return '↑';
3182                         case 'down':
3183                                 return '↓';
3184                 }
3185         }
3186
3187         /**
3188          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3189          *
3190          * @return bool
3191          */
3192         function linkPrefixExtension() {
3193                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3194         }
3195
3196         /**
3197          * Get all magic words from cache.
3198          * @return array
3199          */
3200         function getMagicWords() {
3201                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3202         }
3203
3204         /**
3205          * Run the LanguageGetMagic hook once.
3206          */
3207         protected function doMagicHook() {
3208                 if ( $this->mMagicHookDone ) {
3209                         return;
3210                 }
3211                 $this->mMagicHookDone = true;
3212                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ], '1.16' );
3213         }
3214
3215         /**
3216          * Fill a MagicWord object with data from here
3217          *
3218          * @param MagicWord $mw
3219          */
3220         function getMagic( $mw ) {
3221                 // Saves a function call
3222                 if ( !$this->mMagicHookDone ) {
3223                         $this->doMagicHook();
3224                 }
3225
3226                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3227                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3228                 } else {
3229                         $rawEntry = self::$dataCache->getSubitem(
3230                                 $this->mCode, 'magicWords', $mw->mId );
3231                 }
3232
3233                 if ( !is_array( $rawEntry ) ) {
3234                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3235                 } else {
3236                         $mw->mCaseSensitive = $rawEntry[0];
3237                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3238                 }
3239         }
3240
3241         /**
3242          * Add magic words to the extension array
3243          *
3244          * @param array $newWords
3245          */
3246         function addMagicWordsByLang( $newWords ) {
3247                 $fallbackChain = $this->getFallbackLanguages();
3248                 $fallbackChain = array_reverse( $fallbackChain );
3249                 foreach ( $fallbackChain as $code ) {
3250                         if ( isset( $newWords[$code] ) ) {
3251                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3252                         }
3253                 }
3254         }
3255
3256         /**
3257          * Get special page names, as an associative array
3258          *   canonical name => array of valid names, including aliases
3259          * @return array
3260          */
3261         function getSpecialPageAliases() {
3262                 // Cache aliases because it may be slow to load them
3263                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3264                         // Initialise array
3265                         $this->mExtendedSpecialPageAliases =
3266                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3267                         Hooks::run( 'LanguageGetSpecialPageAliases',
3268                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ], '1.16' );
3269                 }
3270
3271                 return $this->mExtendedSpecialPageAliases;
3272         }
3273
3274         /**
3275          * Italic is unsuitable for some languages
3276          *
3277          * @param string $text The text to be emphasized.
3278          * @return string
3279          */
3280         function emphasize( $text ) {
3281                 return "<em>$text</em>";
3282         }
3283
3284         /**
3285          * Normally we output all numbers in plain en_US style, that is
3286          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3287          * point twohundredthirtyfive. However this is not suitable for all
3288          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3289          * Icelandic just want to use commas instead of dots, and dots instead
3290          * of commas like "293.291,235".
3291          *
3292          * An example of this function being called:
3293          * <code>
3294          * wfMessage( 'message' )->numParams( $num )->text()
3295          * </code>
3296          *
3297          * See $separatorTransformTable on MessageIs.php for
3298          * the , => . and . => , implementation.
3299          *
3300          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3301          * @param int|float $number The string to be formatted, should be an integer
3302          *   or a floating point number.
3303          * @param bool $nocommafy Set to true for special numbers like dates
3304          * @return string
3305          */
3306         public function formatNum( $number, $nocommafy = false ) {
3307                 global $wgTranslateNumerals;
3308                 if ( !$nocommafy ) {
3309                         $number = $this->commafy( $number );
3310                         $s = $this->separatorTransformTable();
3311                         if ( $s ) {
3312                                 $number = strtr( $number, $s );
3313                         }
3314                 }
3315
3316                 if ( $wgTranslateNumerals ) {
3317                         $s = $this->digitTransformTable();
3318                         if ( $s ) {
3319                                 $number = strtr( $number, $s );
3320                         }
3321                 }
3322
3323                 return (string)$number;
3324         }
3325
3326         /**
3327          * Front-end for non-commafied formatNum
3328          *
3329          * @param int|float $number The string to be formatted, should be an integer
3330          *        or a floating point number.
3331          * @since 1.21
3332          * @return string
3333          */
3334         public function formatNumNoSeparators( $number ) {
3335                 return $this->formatNum( $number, true );
3336         }
3337
3338         /**
3339          * @param string $number
3340          * @return string
3341          */
3342         public function parseFormattedNumber( $number ) {
3343                 $s = $this->digitTransformTable();
3344                 if ( $s ) {
3345                         // eliminate empty array values such as ''. (T66347)
3346                         $s = array_filter( $s );
3347                         $number = strtr( $number, array_flip( $s ) );
3348                 }
3349
3350                 $s = $this->separatorTransformTable();
3351                 if ( $s ) {
3352                         // eliminate empty array values such as ''. (T66347)
3353                         $s = array_filter( $s );
3354                         $number = strtr( $number, array_flip( $s ) );
3355                 }
3356
3357                 $number = strtr( $number, [ ',' => '' ] );
3358                 return $number;
3359         }
3360
3361         /**
3362          * Adds commas to a given number
3363          * @since 1.19
3364          * @param mixed $number
3365          * @return string
3366          */
3367         function commafy( $number ) {
3368                 $digitGroupingPattern = $this->digitGroupingPattern();
3369                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3370                 if ( $number === null ) {
3371                         return '';
3372                 }
3373
3374                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3375                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3376                         // In some languages it's conventional not to insert a thousands separator
3377                         // in numbers that are four digits long (1000-9999).
3378                         if ( $minimumGroupingDigits ) {
3379                                 // Number of '#' characters after last comma in the grouping pattern.
3380                                 // The pattern is hardcoded here, but this would vary for different patterns.
3381                                 $primaryGroupingSize = 3;
3382                                 // Maximum length of a number to suppress digit grouping for.
3383                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3384                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3385                                         return $number;
3386                                 }
3387                         }
3388                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3389                 } else {
3390                         // Ref: http://cldr.unicode.org/translation/number-patterns
3391                         $sign = "";
3392                         if ( intval( $number ) < 0 ) {
3393                                 // For negative numbers apply the algorithm like positive number and add sign.
3394                                 $sign = "-";
3395                                 $number = substr( $number, 1 );
3396                         }
3397                         $integerPart = [];
3398                         $decimalPart = [];
3399                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3400                         preg_match( "/\d+/", $number, $integerPart );
3401                         preg_match( "/\.\d*/", $number, $decimalPart );
3402                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3403                         if ( $groupedNumber === $number ) {
3404                                 // the string does not have any number part. Eg: .12345
3405                                 return $sign . $groupedNumber;
3406                         }
3407                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3408                         while ( $start > 0 ) {
3409                                 $match = $matches[0][$numMatches - 1];
3410                                 $matchLen = strlen( $match );
3411                                 $start = $end - $matchLen;
3412                                 if ( $start < 0 ) {
3413                                         $start = 0;
3414                                 }
3415                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3416                                 $end = $start;
3417                                 if ( $numMatches > 1 ) {
3418                                         // use the last pattern for the rest of the number
3419                                         $numMatches--;
3420                                 }
3421                                 if ( $start > 0 ) {
3422                                         $groupedNumber = "," . $groupedNumber;
3423                                 }
3424                         }
3425                         return $sign . $groupedNumber;
3426                 }
3427         }
3428
3429         /**
3430          * @return string
3431          */
3432         function digitGroupingPattern() {
3433                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3434         }
3435
3436         /**
3437          * @return array
3438          */
3439         function digitTransformTable() {
3440                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3441         }
3442
3443         /**
3444          * @return array
3445          */
3446         function separatorTransformTable() {
3447                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3448         }
3449
3450         /**
3451          * @return int|null
3452          */
3453         function minimumGroupingDigits() {
3454                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3455         }
3456
3457         /**
3458          * Take a list of strings and build a locale-friendly comma-separated
3459          * list, using the local comma-separator message.
3460          * The last two strings are chained with an "and".
3461          *
3462          * @param string[] $list
3463          * @return string
3464          */
3465         public function listToText( array $list ) {
3466                 $itemCount = count( $list );
3467                 if ( $itemCount < 1 ) {
3468                         return '';
3469                 }
3470                 $text = array_pop( $list );
3471                 if ( $itemCount > 1 ) {
3472                         $and = $this->msg( 'and' )->escaped();
3473                         $space = $this->msg( 'word-separator' )->escaped();
3474                         $comma = '';
3475                         if ( $itemCount > 2 ) {
3476                                 $comma = $this->msg( 'comma-separator' )->escaped();
3477                         }
3478                         $text = implode( $comma, $list ) . $and . $space . $text;
3479                 }
3480                 return $text;
3481         }
3482
3483         /**
3484          * Take a list of strings and build a locale-friendly comma-separated
3485          * list, using the local comma-separator message.
3486          * @param string[] $list Array of strings to put in a comma list
3487          * @return string
3488          */
3489         function commaList( array $list ) {
3490                 return implode(
3491                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3492                         $list
3493                 );
3494         }
3495
3496         /**
3497          * Take a list of strings and build a locale-friendly semicolon-separated
3498          * list, using the local semicolon-separator message.
3499          * @param string[] $list Array of strings to put in a semicolon list
3500          * @return string
3501          */
3502         function semicolonList( array $list ) {
3503                 return implode(
3504                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3505                         $list
3506                 );
3507         }
3508
3509         /**
3510          * Same as commaList, but separate it with the pipe instead.
3511          * @param string[] $list Array of strings to put in a pipe list
3512          * @return string
3513          */
3514         function pipeList( array $list ) {
3515                 return implode(
3516                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3517                         $list
3518                 );
3519         }
3520
3521         /**
3522          * This method is deprecated since 1.31 and kept as alias for truncateForDatabase, which
3523          * has replaced it. This method provides truncation suitable for DB.
3524          *
3525          * The database offers limited byte lengths for some columns in the database;
3526          * multi-byte character sets mean we need to ensure that only whole characters
3527          * are included, otherwise broken characters can be passed to the user.
3528          *
3529          * @deprecated since 1.31, use truncateForDatabase or truncateForVisual as appropriate.
3530          *
3531          * @param string $string String to truncate
3532          * @param int $length Maximum length (including ellipsis)
3533          * @param string $ellipsis String to append to the truncated text
3534          * @param bool $adjustLength Subtract length of ellipsis from $length.
3535          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3536          * @return string
3537          */
3538         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3539                 wfDeprecated( __METHOD__, '1.31' );
3540                 return $this->truncateForDatabase( $string, $length, $ellipsis, $adjustLength );
3541         }
3542
3543         /**
3544          * Truncate a string to a specified length in bytes, appending an optional
3545          * string (e.g. for ellipsis)
3546          *
3547          * If $length is negative, the string will be truncated from the beginning
3548          *
3549          * @since 1.31
3550          *
3551          * @param string $string String to truncate
3552          * @param int $length Maximum length in bytes
3553          * @param string $ellipsis String to append to the end of truncated text
3554          * @param bool $adjustLength Subtract length of ellipsis from $length
3555          *
3556          * @return string
3557          */
3558         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3559                 return $this->truncateInternal(
3560                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3561                 );
3562         }
3563
3564         /**
3565          * Truncate a string to a specified number of characters, appending an optional
3566          * string (e.g. for ellipsis).
3567          *
3568          * This provides multibyte version of truncate() method of this class, suitable for truncation
3569          * based on number of characters, instead of number of bytes.
3570          *
3571          * If $length is negative, the string will be truncated from the beginning.
3572          *
3573          * @since 1.31
3574          *
3575          * @param string $string String to truncate
3576          * @param int $length Maximum number of characters
3577          * @param string $ellipsis String to append to the end of truncated text
3578          * @param bool $adjustLength Subtract length of ellipsis from $length
3579          *
3580          * @return string
3581          */
3582         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3583                 // Passing encoding to mb_strlen and mb_substr is optional.
3584                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3585                 // explicit specification of encoding is skipped.
3586                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3587                 return $this->truncateInternal(
3588                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3589                 );
3590         }
3591
3592         /**
3593          * Internal method used for truncation. This method abstracts text truncation into
3594          * one common method, allowing users to provide length measurement function and
3595          * function for finding substring.
3596          *
3597          * For usages, see truncateForDatabase and truncateForVisual.
3598          *
3599          * @param string $string String to truncate
3600          * @param int $length Maximum length of final text
3601          * @param string $ellipsis String to append to the end of truncated text
3602          * @param bool $adjustLength Subtract length of ellipsis from $length
3603          * @param callable $measureLength Callable function used for determining the length of text
3604          * @param callable $getSubstring Callable function used for getting the substrings
3605          *
3606          * @return string
3607          */
3608         private function truncateInternal(
3609                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3610         ) {
3611                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3612                         throw new InvalidArgumentException( 'Invalid callback provided' );
3613                 }
3614
3615                 # Check if there is no need to truncate
3616                 if ( $measureLength( $string ) <= abs( $length ) ) {
3617                         return $string; // no need to truncate
3618                 }
3619
3620                 # Use the localized ellipsis character
3621                 if ( $ellipsis == '...' ) {
3622                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3623                 }
3624                 if ( $length == 0 ) {
3625                         return $ellipsis; // convention
3626                 }
3627
3628                 $stringOriginal = $string;
3629                 # If ellipsis length is >= $length then we can't apply $adjustLength
3630                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3631                         $string = $ellipsis; // this can be slightly unexpected
3632                 # Otherwise, truncate and add ellipsis...
3633                 } else {
3634                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3635                         if ( $length > 0 ) {
3636                                 $length -= $ellipsisLength;
3637                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3638                                 $string = $this->removeBadCharLast( $string );
3639                                 $string = rtrim( $string );
3640                                 $string = $string . $ellipsis;
3641                         } else {
3642                                 $length += $ellipsisLength;
3643                                 $string = $getSubstring( $string, $length ); // ...xyz
3644                                 $string = $this->removeBadCharFirst( $string );
3645                                 $string = ltrim( $string );
3646                                 $string = $ellipsis . $string;
3647                         }
3648                 }
3649
3650                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3651                 # This check is *not* redundant if $adjustLength, due to the single case where
3652                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3653                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3654                         return $string;
3655                 } else {
3656                         return $stringOriginal;
3657                 }
3658         }
3659
3660         /**
3661          * Remove bytes that represent an incomplete Unicode character
3662          * at the end of string (e.g. bytes of the char are missing)
3663          *
3664          * @param string $string
3665          * @return string
3666          */
3667         protected function removeBadCharLast( $string ) {
3668                 if ( $string != '' ) {
3669                         $char = ord( $string[strlen( $string ) - 1] );
3670                         $m = [];
3671                         if ( $char >= 0xc0 ) {
3672                                 # We got the first byte only of a multibyte char; remove it.
3673                                 $string = substr( $string, 0, -1 );
3674                         } elseif ( $char >= 0x80 &&
3675                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3676                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3677                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3678                         ) {
3679                                 # We chopped in the middle of a character; remove it
3680                                 $string = $m[1];
3681                         }
3682                 }
3683                 return $string;
3684         }
3685
3686         /**
3687          * Remove bytes that represent an incomplete Unicode character
3688          * at the start of string (e.g. bytes of the char are missing)
3689          *
3690          * @param string $string
3691          * @return string
3692          */
3693         protected function removeBadCharFirst( $string ) {
3694                 if ( $string != '' ) {
3695                         $char = ord( $string[0] );
3696                         if ( $char >= 0x80 && $char < 0xc0 ) {
3697                                 # We chopped in the middle of a character; remove the whole thing
3698                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3699                         }
3700                 }
3701                 return $string;
3702         }
3703
3704         /**
3705          * Truncate a string of valid HTML to a specified length in bytes,
3706          * appending an optional string (e.g. for ellipses), and return valid HTML
3707          *
3708          * This is only intended for styled/linked text, such as HTML with
3709          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3710          * Also, this will not detect things like "display:none" CSS.
3711          *
3712          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3713          *
3714          * @param string $text HTML string to truncate
3715          * @param int $length (zero/positive) Maximum length (including ellipses)
3716          * @param string $ellipsis String to append to the truncated text
3717          * @return string
3718          */
3719         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3720                 # Use the localized ellipsis character
3721                 if ( $ellipsis == '...' ) {
3722                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3723                 }
3724                 # Check if there is clearly no need to truncate
3725                 if ( $length <= 0 ) {
3726                         return $ellipsis; // no text shown, nothing to format (convention)
3727                 } elseif ( strlen( $text ) <= $length ) {
3728                         return $text; // string short enough even *with* HTML (short-circuit)
3729                 }
3730
3731                 $dispLen = 0; // innerHTML legth so far
3732                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3733                 $tagType = 0; // 0-open, 1-close
3734                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3735                 $entityState = 0; // 0-not entity, 1-entity
3736                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3737                 $openTags = []; // open tag stack
3738                 $maybeState = null; // possible truncation state
3739
3740                 $textLen = strlen( $text );
3741                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3742                 for ( $pos = 0; true; ++$pos ) {
3743                         # Consider truncation once the display length has reached the maximim.
3744                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3745                         # Check that we're not in the middle of a bracket/entity...
3746                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3747                                 if ( !$testingEllipsis ) {
3748                                         $testingEllipsis = true;
3749                                         # Save where we are; we will truncate here unless there turn out to
3750                                         # be so few remaining characters that truncation is not necessary.
3751                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3752                                                 $maybeState = [ $ret, $openTags ]; // save state
3753                                         }
3754                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3755                                         # String in fact does need truncation, the truncation point was OK.
3756                                         list( $ret, $openTags ) = $maybeState; // reload state
3757                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3758                                         $ret .= $ellipsis; // add ellipsis
3759                                         break;
3760                                 }
3761                         }
3762                         if ( $pos >= $textLen ) {
3763                                 break; // extra iteration just for above checks
3764                         }
3765
3766                         # Read the next char...
3767                         $ch = $text[$pos];
3768                         $lastCh = $pos ? $text[$pos - 1] : '';
3769                         $ret .= $ch; // add to result string
3770                         if ( $ch == '<' ) {
3771                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3772                                 $entityState = 0; // for bad HTML
3773                                 $bracketState = 1; // tag started (checking for backslash)
3774                         } elseif ( $ch == '>' ) {
3775                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3776                                 $entityState = 0; // for bad HTML
3777                                 $bracketState = 0; // out of brackets
3778                         } elseif ( $bracketState == 1 ) {
3779                                 if ( $ch == '/' ) {
3780                                         $tagType = 1; // close tag (e.g. "</span>")
3781                                 } else {
3782                                         $tagType = 0; // open tag (e.g. "<span>")
3783                                         $tag .= $ch;
3784                                 }
3785                                 $bracketState = 2; // building tag name
3786                         } elseif ( $bracketState == 2 ) {
3787                                 if ( $ch != ' ' ) {
3788                                         $tag .= $ch;
3789                                 } else {
3790                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3791                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3792                                 }
3793                         } elseif ( $bracketState == 0 ) {
3794                                 if ( $entityState ) {
3795                                         if ( $ch == ';' ) {
3796                                                 $entityState = 0;
3797                                                 $dispLen++; // entity is one displayed char
3798                                         }
3799                                 } else {
3800                                         if ( $neLength == 0 && !$maybeState ) {
3801                                                 // Save state without $ch. We want to *hit* the first
3802                                                 // display char (to get tags) but not *use* it if truncating.
3803                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3804                                         }
3805                                         if ( $ch == '&' ) {
3806                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3807                                         } else {
3808                                                 $dispLen++; // this char is displayed
3809                                                 // Add the next $max display text chars after this in one swoop...
3810                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3811                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3812                                                 $dispLen += $skipped;
3813                                                 $pos += $skipped;
3814                                         }
3815                                 }
3816                         }
3817                 }
3818                 // Close the last tag if left unclosed by bad HTML
3819                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3820                 while ( count( $openTags ) > 0 ) {
3821                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3822                 }
3823                 return $ret;
3824         }
3825
3826         /**
3827          * truncateHtml() helper function
3828          * like strcspn() but adds the skipped chars to $ret
3829          *
3830          * @param string $ret
3831          * @param string $text
3832          * @param string $search
3833          * @param int $start
3834          * @param null|int $len
3835          * @return int
3836          */
3837         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3838                 if ( $len === null ) {
3839                         $len = -1; // -1 means "no limit" for strcspn
3840                 } elseif ( $len < 0 ) {
3841                         $len = 0; // sanity
3842                 }
3843                 $skipCount = 0;
3844                 if ( $start < strlen( $text ) ) {
3845                         $skipCount = strcspn( $text, $search, $start, $len );
3846                         $ret .= substr( $text, $start, $skipCount );
3847                 }
3848                 return $skipCount;
3849         }
3850
3851         /**
3852          * truncateHtml() helper function
3853          * (a) push or pop $tag from $openTags as needed
3854          * (b) clear $tag value
3855          * @param string &$tag Current HTML tag name we are looking at
3856          * @param int $tagType (0-open tag, 1-close tag)
3857          * @param string $lastCh Character before the '>' that ended this tag
3858          * @param array &$openTags Open tag stack (not accounting for $tag)
3859          */
3860         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3861                 $tag = ltrim( $tag );
3862                 if ( $tag != '' ) {
3863                         if ( $tagType == 0 && $lastCh != '/' ) {
3864                                 $openTags[] = $tag; // tag opened (didn't close itself)
3865                         } elseif ( $tagType == 1 ) {
3866                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3867                                         array_pop( $openTags ); // tag closed
3868                                 }
3869                         }
3870                         $tag = '';
3871                 }
3872         }
3873
3874         /**
3875          * Grammatical transformations, needed for inflected languages
3876          * Invoked by putting {{grammar:case|word}} in a message
3877          *
3878          * @param string $word
3879          * @param string $case
3880          * @return string
3881          */
3882         function convertGrammar( $word, $case ) {
3883                 global $wgGrammarForms;
3884                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3885                         return $wgGrammarForms[$this->getCode()][$case][$word];
3886                 }
3887
3888                 $grammarTransformations = $this->getGrammarTransformations();
3889
3890                 if ( isset( $grammarTransformations[$case] ) ) {
3891                         $forms = $grammarTransformations[$case];
3892
3893                         // Some names of grammar rules are aliases for other rules.
3894                         // In such cases the value is a string rather than object,
3895                         // so load the actual rules.
3896                         if ( is_string( $forms ) ) {
3897                                 $forms = $grammarTransformations[$forms];
3898                         }
3899
3900                         foreach ( array_values( $forms ) as $rule ) {
3901                                 $form = $rule[0];
3902
3903                                 if ( $form === '@metadata' ) {
3904                                         continue;
3905                                 }
3906
3907                                 $replacement = $rule[1];
3908
3909                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3910                                 $patternMatches = preg_match( $regex, $word );
3911
3912                                 if ( $patternMatches === false ) {
3913                                         wfLogWarning(
3914                                                 'An error occurred while processing grammar. ' .
3915                                                 "Word: '$word'. Regex: /$form/."
3916                                         );
3917                                 } elseif ( $patternMatches === 1 ) {
3918                                         $word = preg_replace( $regex, $replacement, $word );
3919
3920                                         break;
3921                                 }
3922                         }
3923                 }
3924
3925                 return $word;
3926         }
3927
3928         /**
3929          * Get the grammar forms for the content language
3930          * @return array Array of grammar forms
3931          * @since 1.20
3932          */
3933         function getGrammarForms() {
3934                 global $wgGrammarForms;
3935                 if ( isset( $wgGrammarForms[$this->getCode()] )
3936                         && is_array( $wgGrammarForms[$this->getCode()] )
3937                 ) {
3938                         return $wgGrammarForms[$this->getCode()];
3939                 }
3940
3941                 return [];
3942         }
3943
3944         /**
3945          * Get the grammar transformations data for the language.
3946          * Used like grammar forms, with {{GRAMMAR}} and cases,
3947          * but uses pairs of regexes and replacements instead of code.
3948          *
3949          * @return array[] Array of grammar transformations.
3950          * @throws MWException
3951          * @since 1.28
3952          */
3953         public function getGrammarTransformations() {
3954                 $languageCode = $this->getCode();
3955
3956                 if ( self::$grammarTransformations === null ) {
3957                         self::$grammarTransformations = new MapCacheLRU( 10 );
3958                 }
3959
3960                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3961                         return self::$grammarTransformations->get( $languageCode );
3962                 }
3963
3964                 $data = [];
3965
3966                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3967                 if ( is_readable( $grammarDataFile ) ) {
3968                         $data = FormatJson::decode(
3969                                 file_get_contents( $grammarDataFile ),
3970                                 true
3971                         );
3972
3973                         if ( $data === null ) {
3974                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3975                         }
3976
3977                         self::$grammarTransformations->set( $languageCode, $data );
3978                 }
3979
3980                 return $data;
3981         }
3982
3983         /**
3984          * Provides an alternative text depending on specified gender.
3985          * Usage {{gender:username|masculine|feminine|unknown}}.
3986          * username is optional, in which case the gender of current user is used,
3987          * but only in (some) interface messages; otherwise default gender is used.
3988          *
3989          * If no forms are given, an empty string is returned. If only one form is
3990          * given, it will be returned unconditionally. These details are implied by
3991          * the caller and cannot be overridden in subclasses.
3992          *
3993          * If three forms are given, the default is to use the third (unknown) form.
3994          * If fewer than three forms are given, the default is to use the first (masculine) form.
3995          * These details can be overridden in subclasses.
3996          *
3997          * @param string $gender
3998          * @param array $forms
3999          *
4000          * @return string
4001          */
4002         function gender( $gender, $forms ) {
4003                 if ( !count( $forms ) ) {
4004                         return '';
4005                 }
4006                 $forms = $this->preConvertPlural( $forms, 2 );
4007                 if ( $gender === 'male' ) {
4008                         return $forms[0];
4009                 }
4010                 if ( $gender === 'female' ) {
4011                         return $forms[1];
4012                 }
4013                 return $forms[2] ?? $forms[0];
4014         }
4015
4016         /**
4017          * Plural form transformations, needed for some languages.
4018          * For example, there are 3 form of plural in Russian and Polish,
4019          * depending on "count mod 10". See [[w:Plural]]
4020          * For English it is pretty simple.
4021          *
4022          * Invoked by putting {{plural:count|wordform1|wordform2}}
4023          * or {{plural:count|wordform1|wordform2|wordform3}}
4024          *
4025          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
4026          *
4027          * @param int $count Non-localized number
4028          * @param array $forms Different plural forms
4029          * @return string Correct form of plural for $count in this language
4030          */
4031         function convertPlural( $count, $forms ) {
4032                 // Handle explicit n=pluralform cases
4033                 $forms = $this->handleExplicitPluralForms( $count, $forms );
4034                 if ( is_string( $forms ) ) {
4035                         return $forms;
4036                 }
4037                 if ( !count( $forms ) ) {
4038                         return '';
4039                 }
4040
4041                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
4042                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4043                 return $forms[$pluralForm];
4044         }
4045
4046         /**
4047          * Handles explicit plural forms for Language::convertPlural()
4048          *
4049          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4050          * If an explicitly defined plural form matches the $count, then
4051          * string value returned, otherwise array returned for further consideration
4052          * by CLDR rules or overridden convertPlural().
4053          *
4054          * @since 1.23
4055          *
4056          * @param int $count Non-localized number
4057          * @param array $forms Different plural forms
4058          *
4059          * @return array|string
4060          */
4061         protected function handleExplicitPluralForms( $count, array $forms ) {
4062                 foreach ( $forms as $index => $form ) {
4063                         if ( preg_match( '/\d+=/i', $form ) ) {
4064                                 $pos = strpos( $form, '=' );
4065                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4066                                         return substr( $form, $pos + 1 );
4067                                 }
4068                                 unset( $forms[$index] );
4069                         }
4070                 }
4071                 return array_values( $forms );
4072         }
4073
4074         /**
4075          * Checks that convertPlural was given an array and pads it to requested
4076          * amount of forms by copying the last one.
4077          *
4078          * @param array $forms Array of forms given to convertPlural
4079          * @param int $count How many forms should there be at least
4080          * @return array Padded array of forms or an exception if not an array
4081          */
4082         protected function preConvertPlural( /* Array */ $forms, $count ) {
4083                 while ( count( $forms ) < $count ) {
4084                         $forms[] = $forms[count( $forms ) - 1];
4085                 }
4086                 return $forms;
4087         }
4088
4089         /**
4090          * Wraps argument with unicode control characters for directionality safety
4091          *
4092          * This solves the problem where directionality-neutral characters at the edge of
4093          * the argument string get interpreted with the wrong directionality from the
4094          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4095          *
4096          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4097          * directionality of the argument string, using the BIDI algorithm's own "First
4098          * strong directional codepoint" rule. Essentially, this works round the fact that
4099          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4100          * direction inference). The latter is cleaner but still not widely supported.
4101          *
4102          * @param string $text Text to wrap
4103          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4104          */
4105         public function embedBidi( $text = '' ) {
4106                 $dir = self::strongDirFromContent( $text );
4107                 if ( $dir === 'ltr' ) {
4108                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4109                         return self::$lre . $text . self::$pdf;
4110                 }
4111                 if ( $dir === 'rtl' ) {
4112                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4113                         return self::$rle . $text . self::$pdf;
4114                 }
4115                 // No strong directionality: do not wrap
4116                 return $text;
4117         }
4118
4119         /**
4120          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4121          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4122          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4123          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4124          * match up with it.
4125          *
4126          * @param string $str The validated block duration in English
4127          * @param User|null $user User object to use timezone from or null for $wgUser
4128          * @param int $now Current timestamp, for formatting relative block durations
4129          * @return string Somehow translated block duration
4130          * @see LanguageFi.php for example implementation
4131          */
4132         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4133                 $duration = SpecialBlock::getSuggestedDurations( $this );
4134                 foreach ( $duration as $show => $value ) {
4135                         if ( strcmp( $str, $value ) == 0 ) {
4136                                 return htmlspecialchars( trim( $show ) );
4137                         }
4138                 }
4139
4140                 if ( wfIsInfinity( $str ) ) {
4141                         foreach ( $duration as $show => $value ) {
4142                                 if ( wfIsInfinity( $value ) ) {
4143                                         return htmlspecialchars( trim( $show ) );
4144                                 }
4145                         }
4146                 }
4147
4148                 // If all else fails, return a standard duration or timestamp description.
4149                 $time = strtotime( $str, $now );
4150                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4151                         return $str;
4152                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4153                         // The result differs based on current time, so the difference
4154                         // is a fixed duration length.
4155                         return $this->formatDuration( $time - $now );
4156                 } else { // It's an absolute timestamp.
4157                         if ( $time === 0 ) {
4158                                 // wfTimestamp() handles 0 as current time instead of epoch.
4159                                 $time = '19700101000000';
4160                         }
4161                         if ( $user ) {
4162                                 return $this->userTimeAndDate( $time, $user );
4163                         }
4164                         return $this->timeanddate( $time );
4165                 }
4166         }
4167
4168         /**
4169          * languages like Chinese need to be segmented in order for the diff
4170          * to be of any use
4171          *
4172          * @param string $text
4173          * @return string
4174          */
4175         public function segmentForDiff( $text ) {
4176                 return $text;
4177         }
4178
4179         /**
4180          * and unsegment to show the result
4181          *
4182          * @param string $text
4183          * @return string
4184          */
4185         public function unsegmentForDiff( $text ) {
4186                 return $text;
4187         }
4188
4189         /**
4190          * Return the LanguageConverter used in the Language
4191          *
4192          * @since 1.19
4193          * @return LanguageConverter
4194          */
4195         public function getConverter() {
4196                 return $this->mConverter;
4197         }
4198
4199         /**
4200          * convert text to a variant
4201          *
4202          * @param string $text text to convert
4203          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4204          *      variant (if logged in), or the project default variant
4205          * @return string the converted string
4206          */
4207         public function autoConvert( $text, $variant = false ) {
4208                 return $this->mConverter->autoConvert( $text, $variant );
4209         }
4210
4211         /**
4212          * convert text to all supported variants
4213          *
4214          * @param string $text
4215          * @return array
4216          */
4217         public function autoConvertToAllVariants( $text ) {
4218                 return $this->mConverter->autoConvertToAllVariants( $text );
4219         }
4220
4221         /**
4222          * convert text to different variants of a language.
4223          *
4224          * @warning Glossary state is maintained between calls. This means
4225          *  if you pass unescaped text to this method it can cause an XSS
4226          *  in later calls to this method, even if the later calls have properly
4227          *  escaped the input. Never feed this method user controlled text that
4228          *  is not properly escaped!
4229          * @param string $text Content that has been already escaped for use in HTML
4230          * @return string HTML
4231          */
4232         public function convert( $text ) {
4233                 return $this->mConverter->convert( $text );
4234         }
4235
4236         /**
4237          * Convert a Title object to a string in the preferred variant
4238          *
4239          * @param Title $title
4240          * @return string
4241          */
4242         public function convertTitle( $title ) {
4243                 return $this->mConverter->convertTitle( $title );
4244         }
4245
4246         /**
4247          * Convert a namespace index to a string in the preferred variant
4248          *
4249          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4250          * @param string|null $variant variant to convert to, or null to use the user's preferred
4251          *      variant (if logged in), or the project default variant
4252          * @return string a string representation of the namespace
4253          */
4254         public function convertNamespace( $ns, $variant = null ) {
4255                 return $this->mConverter->convertNamespace( $ns, $variant );
4256         }
4257
4258         /**
4259          * Check if this is a language with variants
4260          *
4261          * @return bool
4262          */
4263         public function hasVariants() {
4264                 return count( $this->getVariants() ) > 1;
4265         }
4266
4267         /**
4268          * Check if the language has the specific variant
4269          *
4270          * @since 1.19
4271          * @param string $variant
4272          * @return bool
4273          */
4274         public function hasVariant( $variant ) {
4275                 return (bool)$this->mConverter->validateVariant( $variant );
4276         }
4277
4278         /**
4279          * Perform output conversion on a string, and encode for safe HTML output.
4280          * @param string $text Text to be converted
4281          * @param bool $isTitle Whether this conversion is for the article title
4282          * @return string
4283          * @todo this should get integrated somewhere sane
4284          */
4285         public function convertHtml( $text, $isTitle = false ) {
4286                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4287         }
4288
4289         /**
4290          * @param string $key
4291          * @return string
4292          */
4293         public function convertCategoryKey( $key ) {
4294                 return $this->mConverter->convertCategoryKey( $key );
4295         }
4296
4297         /**
4298          * Get the list of variants supported by this language
4299          * see sample implementation in LanguageZh.php
4300          *
4301          * @return string[] An array of language codes
4302          */
4303         public function getVariants() {
4304                 return $this->mConverter->getVariants();
4305         }
4306
4307         /**
4308          * @return string
4309          */
4310         public function getPreferredVariant() {
4311                 return $this->mConverter->getPreferredVariant();
4312         }
4313
4314         /**
4315          * @return string
4316          */
4317         public function getDefaultVariant() {
4318                 return $this->mConverter->getDefaultVariant();
4319         }
4320
4321         /**
4322          * @return string
4323          */
4324         public function getURLVariant() {
4325                 return $this->mConverter->getURLVariant();
4326         }
4327
4328         /**
4329          * If a language supports multiple variants, it is
4330          * possible that non-existing link in one variant
4331          * actually exists in another variant. this function
4332          * tries to find it. See e.g. LanguageZh.php
4333          * The input parameters may be modified upon return
4334          *
4335          * @param string &$link The name of the link
4336          * @param Title &$nt The title object of the link
4337          * @param bool $ignoreOtherCond To disable other conditions when
4338          *   we need to transclude a template or update a category's link
4339          */
4340         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4341                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4342         }
4343
4344         /**
4345          * returns language specific options used by User::getPageRenderHash()
4346          * for example, the preferred language variant
4347          *
4348          * @return string
4349          */
4350         function getExtraHashOptions() {
4351                 return $this->mConverter->getExtraHashOptions();
4352         }
4353
4354         /**
4355          * For languages that support multiple variants, the title of an
4356          * article may be displayed differently in different variants. this
4357          * function returns the apporiate title defined in the body of the article.
4358          *
4359          * @return string
4360          */
4361         public function getParsedTitle() {
4362                 return $this->mConverter->getParsedTitle();
4363         }
4364
4365         /**
4366          * Refresh the cache of conversion tables when
4367          * MediaWiki:Conversiontable* is updated.
4368          *
4369          * @param Title $title The Title of the page being updated
4370          */
4371         public function updateConversionTable( Title $title ) {
4372                 $this->mConverter->updateConversionTable( $title );
4373         }
4374
4375         /**
4376          * Prepare external link text for conversion. When the text is
4377          * a URL, it shouldn't be converted, and it'll be wrapped in
4378          * the "raw" tag (-{R| }-) to prevent conversion.
4379          *
4380          * This function is called "markNoConversion" for historical
4381          * reasons *BUT DIFFERS SIGNIFICANTLY* from
4382          * LanguageConverter::markNoConversion(), with which it is easily
4383          * confused.
4384          *
4385          * @param string $text Text to be used for external link
4386          * @param bool $noParse Wrap it without confirming it's a real URL first
4387          * @return string The tagged text
4388          * @deprecated since 1.32, use LanguageConverter::markNoConversion()
4389          *  instead.
4390          */
4391         public function markNoConversion( $text, $noParse = false ) {
4392                 wfDeprecated( __METHOD__, '1.32' );
4393                 // Excluding protocal-relative URLs may avoid many false positives.
4394                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4395                         return $this->mConverter->markNoConversion( $text );
4396                 } else {
4397                         return $text;
4398                 }
4399         }
4400
4401         /**
4402          * A regular expression to match legal word-trailing characters
4403          * which should be merged onto a link of the form [[foo]]bar.
4404          *
4405          * @return string
4406          */
4407         public function linkTrail() {
4408                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4409         }
4410
4411         /**
4412          * A regular expression character set to match legal word-prefixing
4413          * characters which should be merged onto a link of the form foo[[bar]].
4414          *
4415          * @return string
4416          */
4417         public function linkPrefixCharset() {
4418                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4419         }
4420
4421         /**
4422          * Get the "parent" language which has a converter to convert a "compatible" language
4423          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4424          *
4425          * @return Language|null
4426          * @since 1.22
4427          */
4428         public function getParentLanguage() {
4429                 if ( $this->mParentLanguage !== false ) {
4430                         return $this->mParentLanguage;
4431                 }
4432
4433                 $code = explode( '-', $this->getCode() )[0];
4434                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4435                         $this->mParentLanguage = null;
4436                         return null;
4437                 }
4438                 $lang = self::factory( $code );
4439                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4440                         $this->mParentLanguage = null;
4441                         return null;
4442                 }
4443
4444                 $this->mParentLanguage = $lang;
4445                 return $lang;
4446         }
4447
4448         /**
4449          * Compare with an other language object
4450          *
4451          * @since 1.28
4452          * @param Language $lang
4453          * @return bool
4454          */
4455         public function equals( Language $lang ) {
4456                 return $lang === $this || $lang->getCode() === $this->mCode;
4457         }
4458
4459         /**
4460          * Get the internal language code for this language object
4461          *
4462          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4463          * htmlspecialchars() or similar
4464          *
4465          * @return string
4466          */
4467         public function getCode() {
4468                 return $this->mCode;
4469         }
4470
4471         /**
4472          * Get the code in BCP 47 format which we can use
4473          * inside of html lang="" tags.
4474          *
4475          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4476          * htmlspecialchars() or similar.
4477          *
4478          * @since 1.19
4479          * @return string
4480          */
4481         public function getHtmlCode() {
4482                 if ( is_null( $this->mHtmlCode ) ) {
4483                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4484                 }
4485                 return $this->mHtmlCode;
4486         }
4487
4488         /**
4489          * @param string $code
4490          */
4491         public function setCode( $code ) {
4492                 $this->mCode = $code;
4493                 // Ensure we don't leave incorrect cached data lying around
4494                 $this->mHtmlCode = null;
4495                 $this->mParentLanguage = false;
4496         }
4497
4498         /**
4499          * Get the language code from a file name. Inverse of getFileName()
4500          * @param string $filename $prefix . $languageCode . $suffix
4501          * @param string $prefix Prefix before the language code
4502          * @param string $suffix Suffix after the language code
4503          * @return string Language code, or false if $prefix or $suffix isn't found
4504          */
4505         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4506                 $m = null;
4507                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4508                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4509                 if ( !count( $m ) ) {
4510                         return false;
4511                 }
4512                 return str_replace( '_', '-', strtolower( $m[1] ) );
4513         }
4514
4515         /**
4516          * @param string $code
4517          * @param bool $fallback Whether we're going through language fallback chain
4518          * @return string Name of the language class
4519          */
4520         public static function classFromCode( $code, $fallback = true ) {
4521                 if ( $fallback && $code == 'en' ) {
4522                         return 'Language';
4523                 } else {
4524                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4525                 }
4526         }
4527
4528         /**
4529          * Get the name of a file for a certain language code
4530          * @param string $prefix Prepend this to the filename
4531          * @param string $code Language code
4532          * @param string $suffix Append this to the filename
4533          * @throws MWException
4534          * @return string $prefix . $mangledCode . $suffix
4535          */
4536         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4537                 if ( !self::isValidBuiltInCode( $code ) ) {
4538                         throw new MWException( "Invalid language code \"$code\"" );
4539                 }
4540
4541                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4542         }
4543
4544         /**
4545          * @param string $code
4546          * @return string
4547          */
4548         public static function getMessagesFileName( $code ) {
4549                 global $IP;
4550                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4551                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4552                 return $file;
4553         }
4554
4555         /**
4556          * @param string $code
4557          * @return string
4558          * @throws MWException
4559          * @since 1.23
4560          */
4561         public static function getJsonMessagesFileName( $code ) {
4562                 global $IP;
4563
4564                 if ( !self::isValidBuiltInCode( $code ) ) {
4565                         throw new MWException( "Invalid language code \"$code\"" );
4566                 }
4567
4568                 return "$IP/languages/i18n/$code.json";
4569         }
4570
4571         /**
4572          * Get the first fallback for a given language.
4573          *
4574          * @param string $code
4575          *
4576          * @return bool|string
4577          */
4578         public static function getFallbackFor( $code ) {
4579                 $fallbacks = self::getFallbacksFor( $code );
4580                 if ( $fallbacks ) {
4581                         return $fallbacks[0];
4582                 }
4583                 return false;
4584         }
4585
4586         /**
4587          * Get the ordered list of fallback languages.
4588          *
4589          * @since 1.19
4590          * @param string $code Language code
4591          * @return array Non-empty array, ending in "en"
4592          */
4593         public static function getFallbacksFor( $code ) {
4594                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4595                         return [];
4596                 }
4597                 // For unknown languages, fallbackSequence returns an empty array,
4598                 // hardcode fallback to 'en' in that case.
4599                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4600         }
4601
4602         /**
4603          * Get the ordered list of fallback languages, ending with the fallback
4604          * language chain for the site language.
4605          *
4606          * @since 1.22
4607          * @param string $code Language code
4608          * @return array Array( fallbacks, site fallbacks )
4609          */
4610         public static function getFallbacksIncludingSiteLanguage( $code ) {
4611                 global $wgLanguageCode;
4612
4613                 // Usually, we will only store a tiny number of fallback chains, so we
4614                 // keep them in static memory.
4615                 $cacheKey = "{$code}-{$wgLanguageCode}";
4616
4617                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4618                         $fallbacks = self::getFallbacksFor( $code );
4619
4620                         // Append the site's fallback chain, including the site language itself
4621                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4622                         array_unshift( $siteFallbacks, $wgLanguageCode );
4623
4624                         // Eliminate any languages already included in the chain
4625                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4626
4627                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4628                 }
4629                 return self::$fallbackLanguageCache[$cacheKey];
4630         }
4631
4632         /**
4633          * Get all messages for a given language
4634          * WARNING: this may take a long time. If you just need all message *keys*
4635          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4636          *
4637          * @param string $code
4638          *
4639          * @return array
4640          */
4641         public static function getMessagesFor( $code ) {
4642                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4643         }
4644
4645         /**
4646          * Get a message for a given language
4647          *
4648          * @param string $key
4649          * @param string $code
4650          *
4651          * @return string
4652          */
4653         public static function getMessageFor( $key, $code ) {
4654                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4655         }
4656
4657         /**
4658          * Get all message keys for a given language. This is a faster alternative to
4659          * array_keys( Language::getMessagesFor( $code ) )
4660          *
4661          * @since 1.19
4662          * @param string $code Language code
4663          * @return array Array of message keys (strings)
4664          */
4665         public static function getMessageKeysFor( $code ) {
4666                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4667         }
4668
4669         /**
4670          * @param string $talk
4671          * @return mixed
4672          */
4673         function fixVariableInNamespace( $talk ) {
4674                 if ( strpos( $talk, '$1' ) === false ) {
4675                         return $talk;
4676                 }
4677
4678                 global $wgMetaNamespace;
4679                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4680
4681                 # Allow grammar transformations
4682                 # Allowing full message-style parsing would make simple requests
4683                 # such as action=raw much more expensive than they need to be.
4684                 # This will hopefully cover most cases.
4685                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4686                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4687                 return str_replace( ' ', '_', $talk );
4688         }
4689
4690         /**
4691          * @param string $m
4692          * @return string
4693          */
4694         function replaceGrammarInNamespace( $m ) {
4695                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4696         }
4697
4698         /**
4699          * Decode an expiry (block, protection, etc) which has come from the DB
4700          *
4701          * @param string $expiry Database expiry String
4702          * @param bool|int $format True to process using language functions, or TS_ constant
4703          *     to return the expiry in a given timestamp
4704          * @param string $infinity If $format is not true, use this string for infinite expiry
4705          * @return string
4706          * @since 1.18
4707          */
4708         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4709                 static $dbInfinity;
4710                 if ( $dbInfinity === null ) {
4711                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4712                 }
4713
4714                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4715                         return $format === true
4716                                 ? $this->getMessageFromDB( 'infiniteblock' )
4717                                 : $infinity;
4718                 } else {
4719                         return $format === true
4720                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4721                                 : wfTimestamp( $format, $expiry );
4722                 }
4723         }
4724
4725         /**
4726          * Formats a time given in seconds into a string representation of that time.
4727          *
4728          * @param int|float $seconds
4729          * @param array $format An optional argument that formats the returned string in different ways:
4730          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4731          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4732          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4733          *     and friends.
4734          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4735          *     or 'avoidminutes'.
4736          * @return string
4737          */
4738         function formatTimePeriod( $seconds, $format = [] ) {
4739                 if ( !is_array( $format ) ) {
4740                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4741                 }
4742                 if ( !isset( $format['avoid'] ) ) {
4743                         $format['avoid'] = false;
4744                 }
4745                 if ( !isset( $format['noabbrevs'] ) ) {
4746                         $format['noabbrevs'] = false;
4747                 }
4748                 $secondsMsg = wfMessage(
4749                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4750                 $minutesMsg = wfMessage(
4751                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4752                 $hoursMsg = wfMessage(
4753                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4754                 $daysMsg = wfMessage(
4755                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4756
4757                 if ( round( $seconds * 10 ) < 100 ) {
4758                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4759                         $s = $secondsMsg->params( $s )->text();
4760                 } elseif ( round( $seconds ) < 60 ) {
4761                         $s = $this->formatNum( round( $seconds ) );
4762                         $s = $secondsMsg->params( $s )->text();
4763                 } elseif ( round( $seconds ) < 3600 ) {
4764                         $minutes = floor( $seconds / 60 );
4765                         $secondsPart = round( fmod( $seconds, 60 ) );
4766                         if ( $secondsPart == 60 ) {
4767                                 $secondsPart = 0;
4768                                 $minutes++;
4769                         }
4770                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4771                         $s .= ' ';
4772                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4773                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4774                         $hours = floor( $seconds / 3600 );
4775                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4776                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4777                         if ( $secondsPart == 60 ) {
4778                                 $secondsPart = 0;
4779                                 $minutes++;
4780                         }
4781                         if ( $minutes == 60 ) {
4782                                 $minutes = 0;
4783                                 $hours++;
4784                         }
4785                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4786                         $s .= ' ';
4787                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4788                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4789                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4790                         }
4791                 } else {
4792                         $days = floor( $seconds / 86400 );
4793                         if ( $format['avoid'] === 'avoidminutes' ) {
4794                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4795                                 if ( $hours == 24 ) {
4796                                         $hours = 0;
4797                                         $days++;
4798                                 }
4799                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4800                                 $s .= ' ';
4801                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4802                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4803                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4804                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4805                                 if ( $minutes == 60 ) {
4806                                         $minutes = 0;
4807                                         $hours++;
4808                                 }
4809                                 if ( $hours == 24 ) {
4810                                         $hours = 0;
4811                                         $days++;
4812                                 }
4813                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4814                                 $s .= ' ';
4815                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4816                                 $s .= ' ';
4817                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4818                         } else {
4819                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4820                                 $s .= ' ';
4821                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4822                         }
4823                 }
4824                 return $s;
4825         }
4826
4827         /**
4828          * Format a bitrate for output, using an appropriate
4829          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4830          *   the magnitude in question.
4831          *
4832          * This use base 1000. For base 1024 use formatSize(), for another base
4833          * see formatComputingNumbers().
4834          *
4835          * @param int $bps
4836          * @return string
4837          */
4838         function formatBitrate( $bps ) {
4839                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4840         }
4841
4842         /**
4843          * @param int $size Size of the unit
4844          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4845          * @param string $messageKey Message key to be uesd
4846          * @return string
4847          */
4848         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4849                 if ( $size <= 0 ) {
4850                         return str_replace( '$1', $this->formatNum( $size ),
4851                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4852                         );
4853                 }
4854                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4855                 $index = 0;
4856
4857                 $maxIndex = count( $sizes ) - 1;
4858                 while ( $size >= $boundary && $index < $maxIndex ) {
4859                         $index++;
4860                         $size /= $boundary;
4861                 }
4862
4863                 // For small sizes no decimal places necessary
4864                 $round = 0;
4865                 if ( $index > 1 ) {
4866                         // For MB and bigger two decimal places are smarter
4867                         $round = 2;
4868                 }
4869                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4870
4871                 $size = round( $size, $round );
4872                 $text = $this->getMessageFromDB( $msg );
4873                 return str_replace( '$1', $this->formatNum( $size ), $text );
4874         }
4875
4876         /**
4877          * Format a size in bytes for output, using an appropriate
4878          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4879          *
4880          * This method use base 1024. For base 1000 use formatBitrate(), for
4881          * another base see formatComputingNumbers()
4882          *
4883          * @param int $size Size to format
4884          * @return string Plain text (not HTML)
4885          */
4886         function formatSize( $size ) {
4887                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4888         }
4889
4890         /**
4891          * Make a list item, used by various special pages
4892          *
4893          * @param string $page Page link
4894          * @param string $details HTML safe text between brackets
4895          * @param bool $oppositedm Add the direction mark opposite to your
4896          *   language, to display text properly
4897          * @return string HTML escaped
4898          */
4899         function specialList( $page, $details, $oppositedm = true ) {
4900                 if ( !$details ) {
4901                         return $page;
4902                 }
4903
4904                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4905                 return $page .
4906                         $dirmark .
4907                         $this->msg( 'word-separator' )->escaped() .
4908                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4909         }
4910
4911         /**
4912          * Generate (prev x| next x) (20|50|100...) type links for paging
4913          *
4914          * @param Title $title Title object to link
4915          * @param int $offset
4916          * @param int $limit
4917          * @param array $query Optional URL query parameter string
4918          * @param bool $atend Optional param for specified if this is the last page
4919          * @return string
4920          */
4921         public function viewPrevNext( Title $title, $offset, $limit,
4922                 array $query = [], $atend = false
4923         ) {
4924                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4925
4926                 # Make 'previous' link
4927                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4928                 if ( $offset > 0 ) {
4929                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4930                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4931                 } else {
4932                         $plink = htmlspecialchars( $prev );
4933                 }
4934
4935                 # Make 'next' link
4936                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4937                 if ( $atend ) {
4938                         $nlink = htmlspecialchars( $next );
4939                 } else {
4940                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4941                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4942                 }
4943
4944                 # Make links to set number of items per page
4945                 $numLinks = [];
4946                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4947                         $numLinks[] = $this->numLink( $title, $offset, $num,
4948                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4949                 }
4950
4951                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4952                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4953         }
4954
4955         /**
4956          * Helper function for viewPrevNext() that generates links
4957          *
4958          * @param Title $title Title object to link
4959          * @param int $offset
4960          * @param int $limit
4961          * @param array $query Extra query parameters
4962          * @param string $link Text to use for the link; will be escaped
4963          * @param string $tooltipMsg Name of the message to use as tooltip
4964          * @param string $class Value of the "class" attribute of the link
4965          * @return string HTML fragment
4966          */
4967         private function numLink( Title $title, $offset, $limit, array $query, $link,
4968                 $tooltipMsg, $class
4969         ) {
4970                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4971                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4972                         ->numParams( $limit )->text();
4973
4974                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4975                         'title' => $tooltip, 'class' => $class ], $link );
4976         }
4977
4978         /**
4979          * Get the conversion rule title, if any.
4980          *
4981          * @return string
4982          */
4983         public function getConvRuleTitle() {
4984                 return $this->mConverter->getConvRuleTitle();
4985         }
4986
4987         /**
4988          * Get the compiled plural rules for the language
4989          * @since 1.20
4990          * @return array Associative array with plural form, and plural rule as key-value pairs
4991          */
4992         public function getCompiledPluralRules() {
4993                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4994                 $fallbacks = self::getFallbacksFor( $this->mCode );
4995                 if ( !$pluralRules ) {
4996                         foreach ( $fallbacks as $fallbackCode ) {
4997                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4998                                 if ( $pluralRules ) {
4999                                         break;
5000                                 }
5001                         }
5002                 }
5003                 return $pluralRules;
5004         }
5005
5006         /**
5007          * Get the plural rules for the language
5008          * @since 1.20
5009          * @return array Associative array with plural form number and plural rule as key-value pairs
5010          */
5011         public function getPluralRules() {
5012                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
5013                 $fallbacks = self::getFallbacksFor( $this->mCode );
5014                 if ( !$pluralRules ) {
5015                         foreach ( $fallbacks as $fallbackCode ) {
5016                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
5017                                 if ( $pluralRules ) {
5018                                         break;
5019                                 }
5020                         }
5021                 }
5022                 return $pluralRules;
5023         }
5024
5025         /**
5026          * Get the plural rule types for the language
5027          * @since 1.22
5028          * @return array Associative array with plural form number and plural rule type as key-value pairs
5029          */
5030         public function getPluralRuleTypes() {
5031                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
5032                 $fallbacks = self::getFallbacksFor( $this->mCode );
5033                 if ( !$pluralRuleTypes ) {
5034                         foreach ( $fallbacks as $fallbackCode ) {
5035                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
5036                                 if ( $pluralRuleTypes ) {
5037                                         break;
5038                                 }
5039                         }
5040                 }
5041                 return $pluralRuleTypes;
5042         }
5043
5044         /**
5045          * Find the index number of the plural rule appropriate for the given number
5046          * @param int $number
5047          * @return int The index number of the plural rule
5048          */
5049         public function getPluralRuleIndexNumber( $number ) {
5050                 $pluralRules = $this->getCompiledPluralRules();
5051                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5052                 return $form;
5053         }
5054
5055         /**
5056          * Find the plural rule type appropriate for the given number
5057          * For example, if the language is set to Arabic, getPluralType(5) should
5058          * return 'few'.
5059          * @since 1.22
5060          * @param int $number
5061          * @return string The name of the plural rule type, e.g. one, two, few, many
5062          */
5063         public function getPluralRuleType( $number ) {
5064                 $index = $this->getPluralRuleIndexNumber( $number );
5065                 $pluralRuleTypes = $this->getPluralRuleTypes();
5066                 if ( isset( $pluralRuleTypes[$index] ) ) {
5067                         return $pluralRuleTypes[$index];
5068                 } else {
5069                         return 'other';
5070                 }
5071         }
5072 }