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