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