languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30
  31 /**
  32  * Internationalisation code
  33  * @ingroup Language
  34  */
  35 class Language {
  36         /**
  37          * Return autonyms in fetchLanguageName(s).
  38          * @since 1.32
  39          */
  40         const AS_AUTONYMS = null;
  41
  42         /**
  43          * Return all known languages in fetchLanguageName(s).
  44          * @since 1.32
  45          */
  46         const ALL = 'all';
  47
  48         /**
  49          * Return in fetchLanguageName(s) only the languages for which we have at
  50          * least some localisation.
  51          * @since 1.32
  52          */
  53         const SUPPORTED = 'mwfile';
  54
  55         /**
  56          * @var LanguageConverter
  57          */
  58         public $mConverter;
  59
  60         public $mVariants, $mCode, $mLoaded = false;
  61         public $mMagicExtensions = [], $mMagicHookDone = false;
  62         private $mHtmlCode = null, $mParentLanguage = false;
  63
  64         public $dateFormatStrings = [];
  65         public $mExtendedSpecialPageAliases;
  66
  67         /** @var array|null */
  68         protected $namespaceNames;
  69         protected $mNamespaceIds, $namespaceAliases;
  70
  71         /**
  72          * ReplacementArray object caches
  73          */
  74         public $transformData = [];
  75
  76         /**
  77          * @var LocalisationCache
  78          */
  79         static public $dataCache;
  80
  81         static public $mLangObjCache = [];
  82
  83         /**
  84          * Return a fallback chain for messages in getFallbacksFor
  85          * @since 1.32
  86          */
  87         const MESSAGES_FALLBACKS = 0;
  88
  89         /**
  90          * Return a strict fallback chain in getFallbacksFor
  91          * @since 1.32
  92          */
  93         const STRICT_FALLBACKS = 1;
  94
  95         static public $mWeekdayMsgs = [
  96                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  97                 'friday', 'saturday'
  98         ];
  99
 100         static public $mWeekdayAbbrevMsgs = [
 101                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 102         ];
 103
 104         static public $mMonthMsgs = [
 105                 'january', 'february', 'march', 'april', 'may_long', 'june',
 106                 'july', 'august', 'september', 'october', 'november',
 107                 'december'
 108         ];
 109         static public $mMonthGenMsgs = [
 110                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 111                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 112                 'december-gen'
 113         ];
 114         static public $mMonthAbbrevMsgs = [
 115                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 116                 'sep', 'oct', 'nov', 'dec'
 117         ];
 118
 119         static public $mIranianCalendarMonthMsgs = [
 120                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 121                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 122                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 123                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 124         ];
 125
 126         static public $mHebrewCalendarMonthMsgs = [
 127                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 128                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 129                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 130                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 131                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 132         ];
 133
 134         static public $mHebrewCalendarMonthGenMsgs = [
 135                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 136                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 137                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 138                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 139                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 140         ];
 141
 142         static public $mHijriCalendarMonthMsgs = [
 143                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 144                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 145                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 146                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 147         ];
 148
 149         /**
 150          * @since 1.20
 151          * @var array
 152          */
 153         static public $durationIntervals = [
 154                 'millennia' => 31556952000,
 155                 'centuries' => 3155695200,
 156                 'decades' => 315569520,
 157                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 158                 'weeks' => 604800,
 159                 'days' => 86400,
 160                 'hours' => 3600,
 161                 'minutes' => 60,
 162                 'seconds' => 1,
 163         ];
 164
 165         /**
 166          * Cache for language fallbacks.
 167          * @see Language::getFallbacksIncludingSiteLanguage
 168          * @since 1.21
 169          * @var array
 170          */
 171         static private $fallbackLanguageCache = [];
 172
 173         /**
 174          * Cache for grammar rules data
 175          * @var MapCacheLRU|null
 176          */
 177         static private $grammarTransformations;
 178
 179         /**
 180          * Cache for language names
 181          * @var HashBagOStuff|null
 182          */
 183         static private $languageNameCache;
 184
 185         /**
 186          * Unicode directional formatting characters, for embedBidi()
 187          */
 188         static private $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 189         static private $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 190         static private $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 191
 192         /**
 193          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 194          * - in group 1 if it is LTR
 195          * - in group 2 if it is RTL
 196          * Does not match if there is no strong directionality codepoint.
 197          *
 198          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 199          *
 200          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 201          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 202          */
 203         // @codeCoverageIgnoreStart
 204         // phpcs:ignore Generic.Files.LineLength
 205         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';
 206         // @codeCoverageIgnoreEnd
 207
 208         /**
 209          * Get a cached or new language object for a given language code
 210          * @param string $code
 211          * @throws MWException
 212          * @return Language
 213          */
 214         static function factory( $code ) {
 215                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 216
 217                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 218                         $code = $wgDummyLanguageCodes[$code];
 219                 }
 220
 221                 // get the language object to process
 222                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 223
 224                 // merge the language object in to get it up front in the cache
 225                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 226                 // get rid of the oldest ones in case we have an overflow
 227                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 228
 229                 return $langObj;
 230         }
 231
 232         /**
 233          * Create a language object for a given language code
 234          * @param string $code
 235          * @param bool $fallback Whether we're going through language fallback chain
 236          * @throws MWException
 237          * @return Language
 238          */
 239         protected static function newFromCode( $code, $fallback = false ) {
 240                 if ( !self::isValidCode( $code ) ) {
 241                         throw new MWException( "Invalid language code \"$code\"" );
 242                 }
 243
 244                 if ( !self::isValidBuiltInCode( $code ) ) {
 245                         // It's not possible to customise this code with class files, so
 246                         // just return a Language object. This is to support uselang= hacks.
 247                         $lang = new Language;
 248                         $lang->setCode( $code );
 249                         return $lang;
 250                 }
 251
 252                 // Check if there is a language class for the code
 253                 $class = self::classFromCode( $code, $fallback );
 254                 // LanguageCode does not inherit Language
 255                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 256                         $lang = new $class;
 257                         return $lang;
 258                 }
 259
 260                 // Keep trying the fallback list until we find an existing class
 261                 $fallbacks = self::getFallbacksFor( $code );
 262                 foreach ( $fallbacks as $fallbackCode ) {
 263                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 264                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 265                         }
 266
 267                         $class = self::classFromCode( $fallbackCode );
 268                         if ( class_exists( $class ) ) {
 269                                 $lang = new $class;
 270                                 $lang->setCode( $code );
 271                                 return $lang;
 272                         }
 273                 }
 274
 275                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 276         }
 277
 278         /**
 279          * Intended for tests that may change configuration in a way that invalidates caches.
 280          *
 281          * @since 1.32
 282          */
 283         public static function clearCaches() {
 284                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 285                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 286                 }
 287                 self::$dataCache = null;
 288                 // Reinitialize $dataCache, since it's expected to always be available
 289                 self::getLocalisationCache();
 290                 self::$mLangObjCache = [];
 291                 self::$fallbackLanguageCache = [];
 292                 self::$grammarTransformations = null;
 293                 self::$languageNameCache = null;
 294         }
 295
 296         /**
 297          * Checks whether any localisation is available for that language tag
 298          * in MediaWiki (MessagesXx.php exists).
 299          *
 300          * @param string $code Language tag (in lower case)
 301          * @return bool Whether language is supported
 302          * @since 1.21
 303          */
 304         public static function isSupportedLanguage( $code ) {
 305                 if ( !self::isValidBuiltInCode( $code ) ) {
 306                         return false;
 307                 }
 308
 309                 if ( $code === 'qqq' ) {
 310                         return false;
 311                 }
 312
 313                 return is_readable( self::getMessagesFileName( $code ) ) ||
 314                         is_readable( self::getJsonMessagesFileName( $code ) );
 315         }
 316
 317         /**
 318          * Returns true if a language code string is a well-formed language tag
 319          * according to RFC 5646.
 320          * This function only checks well-formedness; it doesn't check that
 321          * language, script or variant codes actually exist in the repositories.
 322          *
 323          * Based on regexes by Mark Davis of the Unicode Consortium:
 324          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 325          *
 326          * @param string $code
 327          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 328          *
 329          * @return bool
 330          * @since 1.21
 331          */
 332         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 333                 $alpha = '[a-z]';
 334                 $digit = '[0-9]';
 335                 $alphanum = '[a-z0-9]';
 336                 $x = 'x'; # private use singleton
 337                 $singleton = '[a-wy-z]'; # other singleton
 338                 $s = $lenient ? '[-_]' : '-';
 339
 340                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 341                 $script = "$alpha{4}"; # ISO 15924
 342                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 343                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 344                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 345                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 346
 347                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 348                 # Since these are limited, this is safe even later changes to the registry --
 349                 # the only oddity is that it might change the type of the tag, and thus
 350                 # the results from the capturing groups.
 351                 # https://www.iana.org/assignments/language-subtag-registry
 352
 353                 $grandfathered = "en{$s}GB{$s}oed"
 354                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 355                         . "|no{$s}(?:bok|nyn)"
 356                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 357                         . "|zh{$s}min{$s}nan";
 358
 359                 $variantList = "$variant(?:$s$variant)*";
 360                 $extensionList = "$extension(?:$s$extension)*";
 361
 362                 $langtag = "(?:($language)"
 363                         . "(?:$s$script)?"
 364                         . "(?:$s$region)?"
 365                         . "(?:$s$variantList)?"
 366                         . "(?:$s$extensionList)?"
 367                         . "(?:$s$privateUse)?)";
 368
 369                 # The final breakdown, with capturing groups for each of these components
 370                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 371
 372                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 373
 374                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 375         }
 376
 377         /**
 378          * Returns true if a language code string is of a valid form, whether or
 379          * not it exists. This includes codes which are used solely for
 380          * customisation via the MediaWiki namespace.
 381          *
 382          * @param string $code
 383          *
 384          * @return bool
 385          */
 386         public static function isValidCode( $code ) {
 387                 static $cache = [];
 388                 if ( !isset( $cache[$code] ) ) {
 389                         // People think language codes are html safe, so enforce it.
 390                         // Ideally we should only allow a-zA-Z0-9-
 391                         // but, .+ and other chars are often used for {{int:}} hacks
 392                         // see bugs T39564, T39587, T38938
 393                         $cache[$code] =
 394                                 // Protect against path traversal
 395                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 396                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 397                 }
 398                 return $cache[$code];
 399         }
 400
 401         /**
 402          * Returns true if a language code is of a valid form for the purposes of
 403          * internal customisation of MediaWiki, via Messages*.php or *.json.
 404          *
 405          * @param string $code
 406          *
 407          * @throws MWException
 408          * @since 1.18
 409          * @return bool
 410          */
 411         public static function isValidBuiltInCode( $code ) {
 412                 if ( !is_string( $code ) ) {
 413                         if ( is_object( $code ) ) {
 414                                 $addmsg = " of class " . get_class( $code );
 415                         } else {
 416                                 $addmsg = '';
 417                         }
 418                         $type = gettype( $code );
 419                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 420                 }
 421
 422                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 423         }
 424
 425         /**
 426          * Returns true if a language code is an IETF tag known to MediaWiki.
 427          *
 428          * @param string $tag
 429          *
 430          * @since 1.21
 431          * @return bool
 432          */
 433         public static function isKnownLanguageTag( $tag ) {
 434                 // Quick escape for invalid input to avoid exceptions down the line
 435                 // when code tries to process tags which are not valid at all.
 436                 if ( !self::isValidBuiltInCode( $tag ) ) {
 437                         return false;
 438                 }
 439
 440                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 441                         || self::fetchLanguageName( $tag, $tag ) !== ''
 442                 ) {
 443                         return true;
 444                 }
 445
 446                 return false;
 447         }
 448
 449         /**
 450          * Get the LocalisationCache instance
 451          *
 452          * @return LocalisationCache
 453          */
 454         public static function getLocalisationCache() {
 455                 if ( is_null( self::$dataCache ) ) {
 456                         global $wgLocalisationCacheConf;
 457                         $class = $wgLocalisationCacheConf['class'];
 458                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 459                 }
 460                 return self::$dataCache;
 461         }
 462
 463         function __construct() {
 464                 $this->mConverter = new FakeConverter( $this );
 465                 // Set the code to the name of the descendant
 466                 if ( static::class === 'Language' ) {
 467                         $this->mCode = 'en';
 468                 } else {
 469                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 470                 }
 471                 self::getLocalisationCache();
 472         }
 473
 474         /**
 475          * Reduce memory usage
 476          */
 477         function __destruct() {
 478                 foreach ( $this as $name => $value ) {
 479                         unset( $this->$name );
 480                 }
 481         }
 482
 483         /**
 484          * Hook which will be called if this is the content language.
 485          * Descendants can use this to register hook functions or modify globals
 486          */
 487         function initContLang() {
 488         }
 489
 490         /**
 491          * @return array
 492          * @since 1.19
 493          */
 494         public function getFallbackLanguages() {
 495                 return self::getFallbacksFor( $this->mCode );
 496         }
 497
 498         /**
 499          * Exports $wgBookstoreListEn
 500          * @return array
 501          */
 502         public function getBookstoreList() {
 503                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 504         }
 505
 506         /**
 507          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 508          * available in localised form, it will be included in English.
 509          *
 510          * @return array
 511          */
 512         public function getNamespaces() {
 513                 if ( is_null( $this->namespaceNames ) ) {
 514                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 515
 516                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 517
 518                         $this->namespaceNames = $wgExtraNamespaces +
 519                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 520                         $this->namespaceNames += $validNamespaces;
 521
 522                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 523                         if ( $wgMetaNamespaceTalk ) {
 524                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 525                         } else {
 526                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 527                                 $this->namespaceNames[NS_PROJECT_TALK] =
 528                                         $this->fixVariableInNamespace( $talk );
 529                         }
 530
 531                         # Sometimes a language will be localised but not actually exist on this wiki.
 532                         foreach ( $this->namespaceNames as $key => $text ) {
 533                                 if ( !isset( $validNamespaces[$key] ) ) {
 534                                         unset( $this->namespaceNames[$key] );
 535                                 }
 536                         }
 537
 538                         # The above mixing may leave namespaces out of canonical order.
 539                         # Re-order by namespace ID number...
 540                         ksort( $this->namespaceNames );
 541
 542                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 543                 }
 544
 545                 return $this->namespaceNames;
 546         }
 547
 548         /**
 549          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 550          * @param array $namespaces Array of namespaces (id => name)
 551          */
 552         public function setNamespaces( array $namespaces ) {
 553                 $this->namespaceNames = $namespaces;
 554                 $this->mNamespaceIds = null;
 555         }
 556
 557         /**
 558          * Resets all of the namespace caches. Mainly used for testing
 559          */
 560         public function resetNamespaces() {
 561                 $this->namespaceNames = null;
 562                 $this->mNamespaceIds = null;
 563                 $this->namespaceAliases = null;
 564         }
 565
 566         /**
 567          * A convenience function that returns getNamespaces() with spaces instead of underscores
 568          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 569          *
 570          * @return array
 571          */
 572         public function getFormattedNamespaces() {
 573                 $ns = $this->getNamespaces();
 574                 foreach ( $ns as $k => $v ) {
 575                         $ns[$k] = strtr( $v, '_', ' ' );
 576                 }
 577                 return $ns;
 578         }
 579
 580         /**
 581          * Get a namespace value by key
 582          *
 583          * <code>
 584          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 585          * echo $mw_ns; // prints 'MediaWiki'
 586          * </code>
 587          *
 588          * @param int $index The array key of the namespace to return
 589          * @return string|bool String if the namespace value exists, otherwise false
 590          */
 591         public function getNsText( $index ) {
 592                 $ns = $this->getNamespaces();
 593                 return $ns[$index] ?? false;
 594         }
 595
 596         /**
 597          * A convenience function that returns the same thing as
 598          * getNsText() except with '_' changed to ' ', useful for
 599          * producing output.
 600          *
 601          * <code>
 602          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 603          * echo $mw_ns; // prints 'MediaWiki talk'
 604          * </code>
 605          *
 606          * @param int $index The array key of the namespace to return
 607          * @return string Namespace name without underscores (empty string if namespace does not exist)
 608          */
 609         public function getFormattedNsText( $index ) {
 610                 $ns = $this->getNsText( $index );
 611                 return strtr( $ns, '_', ' ' );
 612         }
 613
 614         /**
 615          * Returns gender-dependent namespace alias if available.
 616          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 617          * @param int $index Namespace index
 618          * @param string $gender Gender key (male, female... )
 619          * @return string
 620          * @since 1.18
 621          */
 622         public function getGenderNsText( $index, $gender ) {
 623                 global $wgExtraGenderNamespaces;
 624
 625                 $ns = $wgExtraGenderNamespaces +
 626                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 627
 628                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 629         }
 630
 631         /**
 632          * Whether this language uses gender-dependent namespace aliases.
 633          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 634          * @return bool
 635          * @since 1.18
 636          */
 637         public function needsGenderDistinction() {
 638                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 639                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 640                         // $wgExtraGenderNamespaces overrides everything
 641                         return true;
 642                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 643                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 644                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 645                         return false;
 646                 } else {
 647                         // Check what is in i18n files
 648                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 649                         return count( $aliases ) > 0;
 650                 }
 651         }
 652
 653         /**
 654          * Get a namespace key by value, case insensitive.
 655          * Only matches namespace names for the current language, not the
 656          * canonical ones defined in Namespace.php.
 657          *
 658          * @param string $text
 659          * @return int|bool An integer if $text is a valid value otherwise false
 660          */
 661         function getLocalNsIndex( $text ) {
 662                 $lctext = $this->lc( $text );
 663                 $ids = $this->getNamespaceIds();
 664                 return $ids[$lctext] ?? false;
 665         }
 666
 667         /**
 668          * @return array
 669          */
 670         public function getNamespaceAliases() {
 671                 if ( is_null( $this->namespaceAliases ) ) {
 672                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 673                         if ( !$aliases ) {
 674                                 $aliases = [];
 675                         } else {
 676                                 foreach ( $aliases as $name => $index ) {
 677                                         if ( $index === NS_PROJECT_TALK ) {
 678                                                 unset( $aliases[$name] );
 679                                                 $name = $this->fixVariableInNamespace( $name );
 680                                                 $aliases[$name] = $index;
 681                                         }
 682                                 }
 683                         }
 684
 685                         global $wgExtraGenderNamespaces;
 686                         $genders = $wgExtraGenderNamespaces +
 687                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 688                         foreach ( $genders as $index => $forms ) {
 689                                 foreach ( $forms as $alias ) {
 690                                         $aliases[$alias] = $index;
 691                                 }
 692                         }
 693
 694                         # Also add converted namespace names as aliases, to avoid confusion.
 695                         $convertedNames = [];
 696                         foreach ( $this->getVariants() as $variant ) {
 697                                 if ( $variant === $this->mCode ) {
 698                                         continue;
 699                                 }
 700                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 701                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 702                                 }
 703                         }
 704
 705                         $this->namespaceAliases = $aliases + $convertedNames;
 706                 }
 707
 708                 return $this->namespaceAliases;
 709         }
 710
 711         /**
 712          * @return array
 713          */
 714         public function getNamespaceIds() {
 715                 if ( is_null( $this->mNamespaceIds ) ) {
 716                         global $wgNamespaceAliases;
 717                         # Put namespace names and aliases into a hashtable.
 718                         # If this is too slow, then we should arrange it so that it is done
 719                         # before caching. The catch is that at pre-cache time, the above
 720                         # class-specific fixup hasn't been done.
 721                         $this->mNamespaceIds = [];
 722                         foreach ( $this->getNamespaces() as $index => $name ) {
 723                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 724                         }
 725                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 726                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 727                         }
 728                         if ( $wgNamespaceAliases ) {
 729                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 730                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 731                                 }
 732                         }
 733                 }
 734                 return $this->mNamespaceIds;
 735         }
 736
 737         /**
 738          * Get a namespace key by value, case insensitive.  Canonical namespace
 739          * names override custom ones defined for the current language.
 740          *
 741          * @param string $text
 742          * @return int|bool An integer if $text is a valid value otherwise false
 743          */
 744         public function getNsIndex( $text ) {
 745                 $lctext = $this->lc( $text );
 746                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 747                 if ( $ns !== null ) {
 748                         return $ns;
 749                 }
 750                 $ids = $this->getNamespaceIds();
 751                 return $ids[$lctext] ?? false;
 752         }
 753
 754         /**
 755          * short names for language variants used for language conversion links.
 756          *
 757          * @param string $code
 758          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 759          * @return string
 760          */
 761         public function getVariantname( $code, $usemsg = true ) {
 762                 $msg = "variantname-$code";
 763                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 764                         return $this->getMessageFromDB( $msg );
 765                 }
 766                 $name = self::fetchLanguageName( $code );
 767                 if ( $name ) {
 768                         return $name; # if it's defined as a language name, show that
 769                 } else {
 770                         # otherwise, output the language code
 771                         return $code;
 772                 }
 773         }
 774
 775         /**
 776          * @return string[]|bool List of date format preference keys, or false if disabled.
 777          */
 778         public function getDatePreferences() {
 779                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 780         }
 781
 782         /**
 783          * @return array
 784          */
 785         function getDateFormats() {
 786                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 787         }
 788
 789         /**
 790          * @return array|string
 791          */
 792         public function getDefaultDateFormat() {
 793                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 794                 if ( $df === 'dmy or mdy' ) {
 795                         global $wgAmericanDates;
 796                         return $wgAmericanDates ? 'mdy' : 'dmy';
 797                 } else {
 798                         return $df;
 799                 }
 800         }
 801
 802         /**
 803          * @return array
 804          */
 805         public function getDatePreferenceMigrationMap() {
 806                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 807         }
 808
 809         /**
 810          * @param string $image
 811          * @return array|null
 812          */
 813         function getImageFile( $image ) {
 814                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 815         }
 816
 817         /**
 818          * @return array
 819          * @since 1.24
 820          */
 821         public function getImageFiles() {
 822                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 823         }
 824
 825         /**
 826          * @return array
 827          */
 828         public function getExtraUserToggles() {
 829                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 830         }
 831
 832         /**
 833          * @param string $tog
 834          * @return string
 835          */
 836         function getUserToggle( $tog ) {
 837                 return $this->getMessageFromDB( "tog-$tog" );
 838         }
 839
 840         /**
 841          * Get an array of language names, indexed by code.
 842          * @param null|string $inLanguage Code of language in which to return the names
 843          *              Use self::AS_AUTONYMS for autonyms (native names)
 844          * @param string $include One of:
 845          *              self::ALL all available languages
 846          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 847          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 848          * @return array Language code => language name (sorted by key)
 849          * @since 1.20
 850          */
 851         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 852                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 853                 $cacheKey .= ":$include";
 854                 if ( self::$languageNameCache === null ) {
 855                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 856                 }
 857
 858                 $ret = self::$languageNameCache->get( $cacheKey );
 859                 if ( !$ret ) {
 860                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 861                         self::$languageNameCache->set( $cacheKey, $ret );
 862                 }
 863                 return $ret;
 864         }
 865
 866         /**
 867          * Uncached helper for fetchLanguageNames
 868          * @param null|string $inLanguage Code of language in which to return the names
 869          *              Use self::AS_AUTONYMS for autonyms (native names)
 870          * @param string $include One of:
 871          *              self::ALL all available languages
 872          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 873          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 874          * @return array Language code => language name (sorted by key)
 875          */
 876         private static function fetchLanguageNamesUncached(
 877                 $inLanguage = self::AS_AUTONYMS,
 878                 $include = 'mw'
 879         ) {
 880                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 881
 882                 // If passed an invalid language code to use, fallback to en
 883                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 884                         $inLanguage = 'en';
 885                 }
 886
 887                 $names = [];
 888
 889                 if ( $inLanguage ) {
 890                         # TODO: also include when $inLanguage is null, when this code is more efficient
 891                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 892                 }
 893
 894                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 895                 if ( $wgUsePigLatinVariant ) {
 896                         // Pig Latin (for variant development)
 897                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 898                 }
 899
 900                 foreach ( $mwNames as $mwCode => $mwName ) {
 901                         # - Prefer own MediaWiki native name when not using the hook
 902                         # - For other names just add if not added through the hook
 903                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 904                                 $names[$mwCode] = $mwName;
 905                         }
 906                 }
 907
 908                 if ( $include === self::ALL ) {
 909                         ksort( $names );
 910                         return $names;
 911                 }
 912
 913                 $returnMw = [];
 914                 $coreCodes = array_keys( $mwNames );
 915                 foreach ( $coreCodes as $coreCode ) {
 916                         $returnMw[$coreCode] = $names[$coreCode];
 917                 }
 918
 919                 if ( $include === self::SUPPORTED ) {
 920                         $namesMwFile = [];
 921                         # We do this using a foreach over the codes instead of a directory
 922                         # loop so that messages files in extensions will work correctly.
 923                         foreach ( $returnMw as $code => $value ) {
 924                                 if ( is_readable( self::getMessagesFileName( $code ) )
 925                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 926                                 ) {
 927                                         $namesMwFile[$code] = $names[$code];
 928                                 }
 929                         }
 930
 931                         ksort( $namesMwFile );
 932                         return $namesMwFile;
 933                 }
 934
 935                 ksort( $returnMw );
 936                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 937                 return $returnMw;
 938         }
 939
 940         /**
 941          * @param string $code The code of the language for which to get the name
 942          * @param null|string $inLanguage Code of language in which to return the name
 943          *   (SELF::AS_AUTONYMS for autonyms)
 944          * @param string $include See fetchLanguageNames()
 945          * @return string Language name or empty
 946          * @since 1.20
 947          */
 948         public static function fetchLanguageName(
 949                 $code,
 950                 $inLanguage = self::AS_AUTONYMS,
 951                 $include = self::ALL
 952         ) {
 953                 $code = strtolower( $code );
 954                 $array = self::fetchLanguageNames( $inLanguage, $include );
 955                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 956         }
 957
 958         /**
 959          * Get a message from the MediaWiki namespace.
 960          *
 961          * @param string $msg Message name
 962          * @return string
 963          */
 964         public function getMessageFromDB( $msg ) {
 965                 return $this->msg( $msg )->text();
 966         }
 967
 968         /**
 969          * Get message object in this language. Only for use inside this class.
 970          *
 971          * @param string $msg Message name
 972          * @return Message
 973          */
 974         protected function msg( $msg ) {
 975                 return wfMessage( $msg )->inLanguage( $this );
 976         }
 977
 978         /**
 979          * @param string $key
 980          * @return string
 981          */
 982         public function getMonthName( $key ) {
 983                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 984         }
 985
 986         /**
 987          * @return array
 988          */
 989         public function getMonthNamesArray() {
 990                 $monthNames = [ '' ];
 991                 for ( $i = 1; $i < 13; $i++ ) {
 992                         $monthNames[] = $this->getMonthName( $i );
 993                 }
 994                 return $monthNames;
 995         }
 996
 997         /**
 998          * @param string $key
 999          * @return string
1000          */
1001         public function getMonthNameGen( $key ) {
1002                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
1003         }
1004
1005         /**
1006          * @param string $key
1007          * @return string
1008          */
1009         public function getMonthAbbreviation( $key ) {
1010                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1011         }
1012
1013         /**
1014          * @return array
1015          */
1016         public function getMonthAbbreviationsArray() {
1017                 $monthNames = [ '' ];
1018                 for ( $i = 1; $i < 13; $i++ ) {
1019                         $monthNames[] = $this->getMonthAbbreviation( $i );
1020                 }
1021                 return $monthNames;
1022         }
1023
1024         /**
1025          * @param string $key
1026          * @return string
1027          */
1028         public function getWeekdayName( $key ) {
1029                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1030         }
1031
1032         /**
1033          * @param string $key
1034          * @return string
1035          */
1036         function getWeekdayAbbreviation( $key ) {
1037                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1038         }
1039
1040         /**
1041          * @param string $key
1042          * @return string
1043          */
1044         function getIranianCalendarMonthName( $key ) {
1045                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1046         }
1047
1048         /**
1049          * @param string $key
1050          * @return string
1051          */
1052         function getHebrewCalendarMonthName( $key ) {
1053                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1054         }
1055
1056         /**
1057          * @param string $key
1058          * @return string
1059          */
1060         function getHebrewCalendarMonthNameGen( $key ) {
1061                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1062         }
1063
1064         /**
1065          * @param string $key
1066          * @return string
1067          */
1068         function getHijriCalendarMonthName( $key ) {
1069                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1070         }
1071
1072         /**
1073          * Pass through result from $dateTimeObj->format()
1074          * @param DateTime|bool|null &$dateTimeObj
1075          * @param string $ts
1076          * @param DateTimeZone|bool|null $zone
1077          * @param string $code
1078          * @return string
1079          */
1080         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1081                 if ( !$dateTimeObj ) {
1082                         $dateTimeObj = DateTime::createFromFormat(
1083                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1084                         );
1085                 }
1086                 return $dateTimeObj->format( $code );
1087         }
1088
1089         /**
1090          * This is a workalike of PHP's date() function, but with better
1091          * internationalisation, a reduced set of format characters, and a better
1092          * escaping format.
1093          *
1094          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1095          * the PHP manual for definitions. There are a number of extensions, which
1096          * start with "x":
1097          *
1098          *    xn   Do not translate digits of the next numeric format character
1099          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1100          *    xr   Use roman numerals for the next numeric format character
1101          *    xh   Use hebrew numerals for the next numeric format character
1102          *    xx   Literal x
1103          *    xg   Genitive month name
1104          *
1105          *    xij  j (day number) in Iranian calendar
1106          *    xiF  F (month name) in Iranian calendar
1107          *    xin  n (month number) in Iranian calendar
1108          *    xiy  y (two digit year) in Iranian calendar
1109          *    xiY  Y (full year) in Iranian calendar
1110          *    xit  t (days in month) in Iranian calendar
1111          *    xiz  z (day of the year) in Iranian calendar
1112          *
1113          *    xjj  j (day number) in Hebrew calendar
1114          *    xjF  F (month name) in Hebrew calendar
1115          *    xjt  t (days in month) in Hebrew calendar
1116          *    xjx  xg (genitive month name) in Hebrew calendar
1117          *    xjn  n (month number) in Hebrew calendar
1118          *    xjY  Y (full year) in Hebrew calendar
1119          *
1120          *    xmj  j (day number) in Hijri calendar
1121          *    xmF  F (month name) in Hijri calendar
1122          *    xmn  n (month number) in Hijri calendar
1123          *    xmY  Y (full year) in Hijri calendar
1124          *
1125          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1126          *                       identical to the Gregorian calendar
1127          *    xoY  Y (full year) in Minguo calendar or Juche year.
1128          *                       Months and days are identical to the
1129          *                       Gregorian calendar
1130          *    xtY  Y (full year) in Japanese nengo. Months and days are
1131          *                       identical to the Gregorian calendar
1132          *
1133          * Characters enclosed in double quotes will be considered literal (with
1134          * the quotes themselves removed). Unmatched quotes will be considered
1135          * literal quotes. Example:
1136          *
1137          * "The month is" F       => The month is January
1138          * i's"                   => 20'11"
1139          *
1140          * Backslash escaping is also supported.
1141          *
1142          * Input timestamp is assumed to be pre-normalized to the desired local
1143          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1144          * $ts is UTC if $zone is not given.
1145          *
1146          * @param string $format
1147          * @param string $ts 14-character timestamp
1148          *      YYYYMMDDHHMMSS
1149          *      01234567890123
1150          * @param DateTimeZone|null $zone Timezone of $ts
1151          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1152          * Only makes sense if $ts is the current time.
1153          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1154          *
1155          * @throws MWException
1156          * @return string
1157          */
1158         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1159                 $s = '';
1160                 $raw = false;
1161                 $roman = false;
1162                 $hebrewNum = false;
1163                 $dateTimeObj = false;
1164                 $rawToggle = false;
1165                 $iranian = false;
1166                 $hebrew = false;
1167                 $hijri = false;
1168                 $thai = false;
1169                 $minguo = false;
1170                 $tenno = false;
1171
1172                 $usedSecond = false;
1173                 $usedMinute = false;
1174                 $usedHour = false;
1175                 $usedAMPM = false;
1176                 $usedDay = false;
1177                 $usedWeek = false;
1178                 $usedMonth = false;
1179                 $usedYear = false;
1180                 $usedISOYear = false;
1181                 $usedIsLeapYear = false;
1182
1183                 $usedHebrewMonth = false;
1184                 $usedIranianMonth = false;
1185                 $usedHijriMonth = false;
1186                 $usedHebrewYear = false;
1187                 $usedIranianYear = false;
1188                 $usedHijriYear = false;
1189                 $usedTennoYear = false;
1190
1191                 if ( strlen( $ts ) !== 14 ) {
1192                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1193                 }
1194
1195                 if ( !ctype_digit( $ts ) ) {
1196                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1197                 }
1198
1199                 $formatLength = strlen( $format );
1200                 for ( $p = 0; $p < $formatLength; $p++ ) {
1201                         $num = false;
1202                         $code = $format[$p];
1203                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1204                                 $code .= $format[++$p];
1205                         }
1206
1207                         if ( ( $code === 'xi'
1208                                         || $code === 'xj'
1209                                         || $code === 'xk'
1210                                         || $code === 'xm'
1211                                         || $code === 'xo'
1212                                         || $code === 'xt' )
1213                                 && $p < $formatLength - 1 ) {
1214                                 $code .= $format[++$p];
1215                         }
1216
1217                         switch ( $code ) {
1218                                 case 'xx':
1219                                         $s .= 'x';
1220                                         break;
1221                                 case 'xn':
1222                                         $raw = true;
1223                                         break;
1224                                 case 'xN':
1225                                         $rawToggle = !$rawToggle;
1226                                         break;
1227                                 case 'xr':
1228                                         $roman = true;
1229                                         break;
1230                                 case 'xh':
1231                                         $hebrewNum = true;
1232                                         break;
1233                                 case 'xg':
1234                                         $usedMonth = true;
1235                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1236                                         break;
1237                                 case 'xjx':
1238                                         $usedHebrewMonth = true;
1239                                         if ( !$hebrew ) {
1240                                                 $hebrew = self::tsToHebrew( $ts );
1241                                         }
1242                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1243                                         break;
1244                                 case 'd':
1245                                         $usedDay = true;
1246                                         $num = substr( $ts, 6, 2 );
1247                                         break;
1248                                 case 'D':
1249                                         $usedDay = true;
1250                                         $s .= $this->getWeekdayAbbreviation(
1251                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1252                                         );
1253                                         break;
1254                                 case 'j':
1255                                         $usedDay = true;
1256                                         $num = intval( substr( $ts, 6, 2 ) );
1257                                         break;
1258                                 case 'xij':
1259                                         $usedDay = true;
1260                                         if ( !$iranian ) {
1261                                                 $iranian = self::tsToIranian( $ts );
1262                                         }
1263                                         $num = $iranian[2];
1264                                         break;
1265                                 case 'xmj':
1266                                         $usedDay = true;
1267                                         if ( !$hijri ) {
1268                                                 $hijri = self::tsToHijri( $ts );
1269                                         }
1270                                         $num = $hijri[2];
1271                                         break;
1272                                 case 'xjj':
1273                                         $usedDay = true;
1274                                         if ( !$hebrew ) {
1275                                                 $hebrew = self::tsToHebrew( $ts );
1276                                         }
1277                                         $num = $hebrew[2];
1278                                         break;
1279                                 case 'l':
1280                                         $usedDay = true;
1281                                         $s .= $this->getWeekdayName(
1282                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1283                                         );
1284                                         break;
1285                                 case 'F':
1286                                         $usedMonth = true;
1287                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1288                                         break;
1289                                 case 'xiF':
1290                                         $usedIranianMonth = true;
1291                                         if ( !$iranian ) {
1292                                                 $iranian = self::tsToIranian( $ts );
1293                                         }
1294                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1295                                         break;
1296                                 case 'xmF':
1297                                         $usedHijriMonth = true;
1298                                         if ( !$hijri ) {
1299                                                 $hijri = self::tsToHijri( $ts );
1300                                         }
1301                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1302                                         break;
1303                                 case 'xjF':
1304                                         $usedHebrewMonth = true;
1305                                         if ( !$hebrew ) {
1306                                                 $hebrew = self::tsToHebrew( $ts );
1307                                         }
1308                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1309                                         break;
1310                                 case 'm':
1311                                         $usedMonth = true;
1312                                         $num = substr( $ts, 4, 2 );
1313                                         break;
1314                                 case 'M':
1315                                         $usedMonth = true;
1316                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1317                                         break;
1318                                 case 'n':
1319                                         $usedMonth = true;
1320                                         $num = intval( substr( $ts, 4, 2 ) );
1321                                         break;
1322                                 case 'xin':
1323                                         $usedIranianMonth = true;
1324                                         if ( !$iranian ) {
1325                                                 $iranian = self::tsToIranian( $ts );
1326                                         }
1327                                         $num = $iranian[1];
1328                                         break;
1329                                 case 'xmn':
1330                                         $usedHijriMonth = true;
1331                                         if ( !$hijri ) {
1332                                                 $hijri = self::tsToHijri( $ts );
1333                                         }
1334                                         $num = $hijri[1];
1335                                         break;
1336                                 case 'xjn':
1337                                         $usedHebrewMonth = true;
1338                                         if ( !$hebrew ) {
1339                                                 $hebrew = self::tsToHebrew( $ts );
1340                                         }
1341                                         $num = $hebrew[1];
1342                                         break;
1343                                 case 'xjt':
1344                                         $usedHebrewMonth = true;
1345                                         if ( !$hebrew ) {
1346                                                 $hebrew = self::tsToHebrew( $ts );
1347                                         }
1348                                         $num = $hebrew[3];
1349                                         break;
1350                                 case 'Y':
1351                                         $usedYear = true;
1352                                         $num = substr( $ts, 0, 4 );
1353                                         break;
1354                                 case 'xiY':
1355                                         $usedIranianYear = true;
1356                                         if ( !$iranian ) {
1357                                                 $iranian = self::tsToIranian( $ts );
1358                                         }
1359                                         $num = $iranian[0];
1360                                         break;
1361                                 case 'xmY':
1362                                         $usedHijriYear = true;
1363                                         if ( !$hijri ) {
1364                                                 $hijri = self::tsToHijri( $ts );
1365                                         }
1366                                         $num = $hijri[0];
1367                                         break;
1368                                 case 'xjY':
1369                                         $usedHebrewYear = true;
1370                                         if ( !$hebrew ) {
1371                                                 $hebrew = self::tsToHebrew( $ts );
1372                                         }
1373                                         $num = $hebrew[0];
1374                                         break;
1375                                 case 'xkY':
1376                                         $usedYear = true;
1377                                         if ( !$thai ) {
1378                                                 $thai = self::tsToYear( $ts, 'thai' );
1379                                         }
1380                                         $num = $thai[0];
1381                                         break;
1382                                 case 'xoY':
1383                                         $usedYear = true;
1384                                         if ( !$minguo ) {
1385                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1386                                         }
1387                                         $num = $minguo[0];
1388                                         break;
1389                                 case 'xtY':
1390                                         $usedTennoYear = true;
1391                                         if ( !$tenno ) {
1392                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1393                                         }
1394                                         $num = $tenno[0];
1395                                         break;
1396                                 case 'y':
1397                                         $usedYear = true;
1398                                         $num = substr( $ts, 2, 2 );
1399                                         break;
1400                                 case 'xiy':
1401                                         $usedIranianYear = true;
1402                                         if ( !$iranian ) {
1403                                                 $iranian = self::tsToIranian( $ts );
1404                                         }
1405                                         $num = substr( $iranian[0], -2 );
1406                                         break;
1407                                 case 'xit':
1408                                         $usedIranianYear = true;
1409                                         if ( !$iranian ) {
1410                                                 $iranian = self::tsToIranian( $ts );
1411                                         }
1412                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1413                                         break;
1414                                 case 'xiz':
1415                                         $usedIranianYear = true;
1416                                         if ( !$iranian ) {
1417                                                 $iranian = self::tsToIranian( $ts );
1418                                         }
1419                                         $num = $iranian[3];
1420                                         break;
1421                                 case 'a':
1422                                         $usedAMPM = true;
1423                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1424                                         break;
1425                                 case 'A':
1426                                         $usedAMPM = true;
1427                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1428                                         break;
1429                                 case 'g':
1430                                         $usedHour = true;
1431                                         $h = substr( $ts, 8, 2 );
1432                                         $num = $h % 12 ? $h % 12 : 12;
1433                                         break;
1434                                 case 'G':
1435                                         $usedHour = true;
1436                                         $num = intval( substr( $ts, 8, 2 ) );
1437                                         break;
1438                                 case 'h':
1439                                         $usedHour = true;
1440                                         $h = substr( $ts, 8, 2 );
1441                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1442                                         break;
1443                                 case 'H':
1444                                         $usedHour = true;
1445                                         $num = substr( $ts, 8, 2 );
1446                                         break;
1447                                 case 'i':
1448                                         $usedMinute = true;
1449                                         $num = substr( $ts, 10, 2 );
1450                                         break;
1451                                 case 's':
1452                                         $usedSecond = true;
1453                                         $num = substr( $ts, 12, 2 );
1454                                         break;
1455                                 case 'c':
1456                                 case 'r':
1457                                         $usedSecond = true;
1458                                         // fall through
1459                                 case 'e':
1460                                 case 'O':
1461                                 case 'P':
1462                                 case 'T':
1463                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1464                                         break;
1465                                 case 'w':
1466                                 case 'N':
1467                                 case 'z':
1468                                         $usedDay = true;
1469                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1470                                         break;
1471                                 case 'W':
1472                                         $usedWeek = true;
1473                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1474                                         break;
1475                                 case 't':
1476                                         $usedMonth = true;
1477                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1478                                         break;
1479                                 case 'L':
1480                                         $usedIsLeapYear = true;
1481                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1482                                         break;
1483                                 case 'o':
1484                                         $usedISOYear = true;
1485                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1486                                         break;
1487                                 case 'U':
1488                                         $usedSecond = true;
1489                                         // fall through
1490                                 case 'I':
1491                                 case 'Z':
1492                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1493                                         break;
1494                                 case '\\':
1495                                         # Backslash escaping
1496                                         if ( $p < $formatLength - 1 ) {
1497                                                 $s .= $format[++$p];
1498                                         } else {
1499                                                 $s .= '\\';
1500                                         }
1501                                         break;
1502                                 case '"':
1503                                         # Quoted literal
1504                                         if ( $p < $formatLength - 1 ) {
1505                                                 $endQuote = strpos( $format, '"', $p + 1 );
1506                                                 if ( $endQuote === false ) {
1507                                                         # No terminating quote, assume literal "
1508                                                         $s .= '"';
1509                                                 } else {
1510                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1511                                                         $p = $endQuote;
1512                                                 }
1513                                         } else {
1514                                                 # Quote at end of string, assume literal "
1515                                                 $s .= '"';
1516                                         }
1517                                         break;
1518                                 default:
1519                                         $s .= $format[$p];
1520                         }
1521                         if ( $num !== false ) {
1522                                 if ( $rawToggle || $raw ) {
1523                                         $s .= $num;
1524                                         $raw = false;
1525                                 } elseif ( $roman ) {
1526                                         $s .= self::romanNumeral( $num );
1527                                         $roman = false;
1528                                 } elseif ( $hebrewNum ) {
1529                                         $s .= self::hebrewNumeral( $num );
1530                                         $hebrewNum = false;
1531                                 } else {
1532                                         $s .= $this->formatNum( $num, true );
1533                                 }
1534                         }
1535                 }
1536
1537                 if ( $ttl === 'unused' ) {
1538                         // No need to calculate the TTL, the caller wont use it anyway.
1539                 } elseif ( $usedSecond ) {
1540                         $ttl = 1;
1541                 } elseif ( $usedMinute ) {
1542                         $ttl = 60 - substr( $ts, 12, 2 );
1543                 } elseif ( $usedHour ) {
1544                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1545                 } elseif ( $usedAMPM ) {
1546                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1547                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1548                 } elseif (
1549                         $usedDay ||
1550                         $usedHebrewMonth ||
1551                         $usedIranianMonth ||
1552                         $usedHijriMonth ||
1553                         $usedHebrewYear ||
1554                         $usedIranianYear ||
1555                         $usedHijriYear ||
1556                         $usedTennoYear
1557                 ) {
1558                         // @todo Someone who understands the non-Gregorian calendars
1559                         // should write proper logic for them so that they don't need purged every day.
1560                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1561                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1562                 } else {
1563                         $possibleTtls = [];
1564                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1565                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1566                         if ( $usedWeek ) {
1567                                 $possibleTtls[] =
1568                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1569                                         $timeRemainingInDay;
1570                         } elseif ( $usedISOYear ) {
1571                                 // December 28th falls on the last ISO week of the year, every year.
1572                                 // The last ISO week of a year can be 52 or 53.
1573                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1574                                         'Ymd',
1575                                         substr( $ts, 0, 4 ) . '1228',
1576                                         $zone ?: new DateTimeZone( 'UTC' )
1577                                 )->format( 'W' );
1578                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1579                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1580                                 $timeRemainingInWeek =
1581                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1582                                         + $timeRemainingInDay;
1583                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1584                         }
1585
1586                         if ( $usedMonth ) {
1587                                 $possibleTtls[] =
1588                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1589                                                 substr( $ts, 6, 2 ) ) * 86400
1590                                         + $timeRemainingInDay;
1591                         } elseif ( $usedYear ) {
1592                                 $possibleTtls[] =
1593                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1594                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1595                                         + $timeRemainingInDay;
1596                         } elseif ( $usedIsLeapYear ) {
1597                                 $year = substr( $ts, 0, 4 );
1598                                 $timeRemainingInYear =
1599                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1600                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1601                                         + $timeRemainingInDay;
1602                                 $mod = $year % 4;
1603                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1604                                         // this isn't a leap year. see when the next one starts
1605                                         $nextCandidate = $year - $mod + 4;
1606                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1607                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1608                                                         $timeRemainingInYear;
1609                                         } else {
1610                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1611                                                         $timeRemainingInYear;
1612                                         }
1613                                 } else {
1614                                         // this is a leap year, so the next year isn't
1615                                         $possibleTtls[] = $timeRemainingInYear;
1616                                 }
1617                         }
1618
1619                         if ( $possibleTtls ) {
1620                                 $ttl = min( $possibleTtls );
1621                         }
1622                 }
1623
1624                 return $s;
1625         }
1626
1627         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1628         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1629
1630         /**
1631          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1632          * Gregorian dates to Iranian dates. Originally written in C, it
1633          * is released under the terms of GNU Lesser General Public
1634          * License. Conversion to PHP was performed by Niklas Laxström.
1635          *
1636          * Link: http://www.farsiweb.info/jalali/jalali.c
1637          *
1638          * @param string $ts
1639          *
1640          * @return int[]
1641          */
1642         private static function tsToIranian( $ts ) {
1643                 $gy = substr( $ts, 0, 4 ) - 1600;
1644                 $gm = substr( $ts, 4, 2 ) - 1;
1645                 $gd = substr( $ts, 6, 2 ) - 1;
1646
1647                 # Days passed from the beginning (including leap years)
1648                 $gDayNo = 365 * $gy
1649                         + floor( ( $gy + 3 ) / 4 )
1650                         - floor( ( $gy + 99 ) / 100 )
1651                         + floor( ( $gy + 399 ) / 400 );
1652
1653                 // Add days of the past months of this year
1654                 for ( $i = 0; $i < $gm; $i++ ) {
1655                         $gDayNo += self::$GREG_DAYS[$i];
1656                 }
1657
1658                 // Leap years
1659                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1660                         $gDayNo++;
1661                 }
1662
1663                 // Days passed in current month
1664                 $gDayNo += (int)$gd;
1665
1666                 $jDayNo = $gDayNo - 79;
1667
1668                 $jNp = floor( $jDayNo / 12053 );
1669                 $jDayNo %= 12053;
1670
1671                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1672                 $jDayNo %= 1461;
1673
1674                 if ( $jDayNo >= 366 ) {
1675                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1676                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1677                 }
1678
1679                 $jz = $jDayNo;
1680
1681                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1682                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1683                 }
1684
1685                 $jm = $i + 1;
1686                 $jd = $jDayNo + 1;
1687
1688                 return [ $jy, $jm, $jd, $jz ];
1689         }
1690
1691         /**
1692          * Converting Gregorian dates to Hijri dates.
1693          *
1694          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1695          *
1696          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1697          *
1698          * @param string $ts
1699          *
1700          * @return int[]
1701          */
1702         private static function tsToHijri( $ts ) {
1703                 $year = substr( $ts, 0, 4 );
1704                 $month = substr( $ts, 4, 2 );
1705                 $day = substr( $ts, 6, 2 );
1706
1707                 $zyr = $year;
1708                 $zd = $day;
1709                 $zm = $month;
1710                 $zy = $zyr;
1711
1712                 if (
1713                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1714                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1715                 ) {
1716                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1717                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1718                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1719                                         $zd - 32075;
1720                 } else {
1721                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1722                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1723                 }
1724
1725                 $zl = $zjd - 1948440 + 10632;
1726                 $zn = (int)( ( $zl - 1 ) / 10631 );
1727                 $zl = $zl - 10631 * $zn + 354;
1728                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1729                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1730                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1731                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1732                 $zm = (int)( ( 24 * $zl ) / 709 );
1733                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1734                 $zy = 30 * $zn + $zj - 30;
1735
1736                 return [ $zy, $zm, $zd ];
1737         }
1738
1739         /**
1740          * Converting Gregorian dates to Hebrew dates.
1741          *
1742          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1743          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1744          * to translate the relevant functions into PHP and release them under
1745          * GNU GPL.
1746          *
1747          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1748          * and Adar II is 14. In a non-leap year, Adar is 6.
1749          *
1750          * @param string $ts
1751          *
1752          * @return int[]
1753          */
1754         private static function tsToHebrew( $ts ) {
1755                 # Parse date
1756                 $year = substr( $ts, 0, 4 );
1757                 $month = substr( $ts, 4, 2 );
1758                 $day = substr( $ts, 6, 2 );
1759
1760                 # Calculate Hebrew year
1761                 $hebrewYear = $year + 3760;
1762
1763                 # Month number when September = 1, August = 12
1764                 $month += 4;
1765                 if ( $month > 12 ) {
1766                         # Next year
1767                         $month -= 12;
1768                         $year++;
1769                         $hebrewYear++;
1770                 }
1771
1772                 # Calculate day of year from 1 September
1773                 $dayOfYear = $day;
1774                 for ( $i = 1; $i < $month; $i++ ) {
1775                         if ( $i == 6 ) {
1776                                 # February
1777                                 $dayOfYear += 28;
1778                                 # Check if the year is leap
1779                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1780                                         $dayOfYear++;
1781                                 }
1782                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1783                                 $dayOfYear += 30;
1784                         } else {
1785                                 $dayOfYear += 31;
1786                         }
1787                 }
1788
1789                 # Calculate the start of the Hebrew year
1790                 $start = self::hebrewYearStart( $hebrewYear );
1791
1792                 # Calculate next year's start
1793                 if ( $dayOfYear <= $start ) {
1794                         # Day is before the start of the year - it is the previous year
1795                         # Next year's start
1796                         $nextStart = $start;
1797                         # Previous year
1798                         $year--;
1799                         $hebrewYear--;
1800                         # Add days since previous year's 1 September
1801                         $dayOfYear += 365;
1802                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1803                                 # Leap year
1804                                 $dayOfYear++;
1805                         }
1806                         # Start of the new (previous) year
1807                         $start = self::hebrewYearStart( $hebrewYear );
1808                 } else {
1809                         # Next year's start
1810                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1811                 }
1812
1813                 # Calculate Hebrew day of year
1814                 $hebrewDayOfYear = $dayOfYear - $start;
1815
1816                 # Difference between year's days
1817                 $diff = $nextStart - $start;
1818                 # Add 12 (or 13 for leap years) days to ignore the difference between
1819                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1820                 # difference is only about the year type
1821                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1822                         $diff += 13;
1823                 } else {
1824                         $diff += 12;
1825                 }
1826
1827                 # Check the year pattern, and is leap year
1828                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1829                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1830                 # and non-leap years
1831                 $yearPattern = $diff % 30;
1832                 # Check if leap year
1833                 $isLeap = $diff >= 30;
1834
1835                 # Calculate day in the month from number of day in the Hebrew year
1836                 # Don't check Adar - if the day is not in Adar, we will stop before;
1837                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1838                 $hebrewDay = $hebrewDayOfYear;
1839                 $hebrewMonth = 1;
1840                 $days = 0;
1841                 while ( $hebrewMonth <= 12 ) {
1842                         # Calculate days in this month
1843                         if ( $isLeap && $hebrewMonth == 6 ) {
1844                                 # Adar in a leap year
1845                                 if ( $isLeap ) {
1846                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1847                                         $days = 30;
1848                                         if ( $hebrewDay <= $days ) {
1849                                                 # Day in Adar I
1850                                                 $hebrewMonth = 13;
1851                                         } else {
1852                                                 # Subtract the days of Adar I
1853                                                 $hebrewDay -= $days;
1854                                                 # Try Adar II
1855                                                 $days = 29;
1856                                                 if ( $hebrewDay <= $days ) {
1857                                                         # Day in Adar II
1858                                                         $hebrewMonth = 14;
1859                                                 }
1860                                         }
1861                                 }
1862                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1863                                 # Cheshvan in a complete year (otherwise as the rule below)
1864                                 $days = 30;
1865                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1866                                 # Kislev in an incomplete year (otherwise as the rule below)
1867                                 $days = 29;
1868                         } else {
1869                                 # Odd months have 30 days, even have 29
1870                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1871                         }
1872                         if ( $hebrewDay <= $days ) {
1873                                 # In the current month
1874                                 break;
1875                         } else {
1876                                 # Subtract the days of the current month
1877                                 $hebrewDay -= $days;
1878                                 # Try in the next month
1879                                 $hebrewMonth++;
1880                         }
1881                 }
1882
1883                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1884         }
1885
1886         /**
1887          * This calculates the Hebrew year start, as days since 1 September.
1888          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1889          * Used for Hebrew date.
1890          *
1891          * @param int $year
1892          *
1893          * @return string
1894          */
1895         private static function hebrewYearStart( $year ) {
1896                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1897                 $b = intval( ( $year - 1 ) % 4 );
1898                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1899                 if ( $m < 0 ) {
1900                         $m--;
1901                 }
1902                 $Mar = intval( $m );
1903                 if ( $m < 0 ) {
1904                         $m++;
1905                 }
1906                 $m -= $Mar;
1907
1908                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1909                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1910                         $Mar++;
1911                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1912                         $Mar += 2;
1913                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1914                         $Mar++;
1915                 }
1916
1917                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1918                 return $Mar;
1919         }
1920
1921         /**
1922          * Algorithm to convert Gregorian dates to Thai solar dates,
1923          * Minguo dates or Minguo dates.
1924          *
1925          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1926          *       https://en.wikipedia.org/wiki/Minguo_calendar
1927          *       https://en.wikipedia.org/wiki/Japanese_era_name
1928          *
1929          * @param string $ts 14-character timestamp
1930          * @param string $cName Calender name
1931          * @return array Converted year, month, day
1932          */
1933         private static function tsToYear( $ts, $cName ) {
1934                 $gy = substr( $ts, 0, 4 );
1935                 $gm = substr( $ts, 4, 2 );
1936                 $gd = substr( $ts, 6, 2 );
1937
1938                 if ( !strcmp( $cName, 'thai' ) ) {
1939                         # Thai solar dates
1940                         # Add 543 years to the Gregorian calendar
1941                         # Months and days are identical
1942                         $gy_offset = $gy + 543;
1943                         # fix for dates between 1912 and 1941
1944                         # https://en.wikipedia.org/?oldid=836596673#New_year
1945                         if ( $gy >= 1912 && $gy <= 1940 ) {
1946                                 if ( $gm <= 3 ) {
1947                                         $gy_offset--;
1948                                 }
1949                                 $gm = ( $gm - 3 ) % 12;
1950                         }
1951                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1952                         # Minguo dates
1953                         # Deduct 1911 years from the Gregorian calendar
1954                         # Months and days are identical
1955                         $gy_offset = $gy - 1911;
1956                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1957                         # Nengō dates up to Meiji period
1958                         # Deduct years from the Gregorian calendar
1959                         # depending on the nengo periods
1960                         # Months and days are identical
1961                         if ( ( $gy < 1912 )
1962                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1963                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1964                         ) {
1965                                 # Meiji period
1966                                 $gy_gannen = $gy - 1868 + 1;
1967                                 $gy_offset = $gy_gannen;
1968                                 if ( $gy_gannen == 1 ) {
1969                                         $gy_offset = '元';
1970                                 }
1971                                 $gy_offset = '明治' . $gy_offset;
1972                         } elseif (
1973                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1974                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1975                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1976                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1977                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1978                         ) {
1979                                 # Taishō period
1980                                 $gy_gannen = $gy - 1912 + 1;
1981                                 $gy_offset = $gy_gannen;
1982                                 if ( $gy_gannen == 1 ) {
1983                                         $gy_offset = '元';
1984                                 }
1985                                 $gy_offset = '大正' . $gy_offset;
1986                         } elseif (
1987                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1988                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1989                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1990                         ) {
1991                                 # Shōwa period
1992                                 $gy_gannen = $gy - 1926 + 1;
1993                                 $gy_offset = $gy_gannen;
1994                                 if ( $gy_gannen == 1 ) {
1995                                         $gy_offset = '元';
1996                                 }
1997                                 $gy_offset = '昭和' . $gy_offset;
1998                         } else {
1999                                 # Heisei period
2000                                 $gy_gannen = $gy - 1989 + 1;
2001                                 $gy_offset = $gy_gannen;
2002                                 if ( $gy_gannen == 1 ) {
2003                                         $gy_offset = '元';
2004                                 }
2005                                 $gy_offset = '平成' . $gy_offset;
2006                         }
2007                 } else {
2008                         $gy_offset = $gy;
2009                 }
2010
2011                 return [ $gy_offset, $gm, $gd ];
2012         }
2013
2014         /**
2015          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2016          *
2017          * This is the rule the BIDI algorithm uses to determine the directionality of
2018          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2019          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2020          *
2021          * TODO: Does not handle BIDI control characters inside the text.
2022          * TODO: Does not handle unallocated characters.
2023          *
2024          * @param string $text Text to test
2025          * @return null|string Directionality ('ltr' or 'rtl') or null
2026          */
2027         private static function strongDirFromContent( $text = '' ) {
2028                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2029                         return null;
2030                 }
2031                 if ( $matches[1] === '' ) {
2032                         return 'rtl';
2033                 }
2034                 return 'ltr';
2035         }
2036
2037         /**
2038          * Roman number formatting up to 10000
2039          *
2040          * @param int $num
2041          *
2042          * @return string
2043          */
2044         static function romanNumeral( $num ) {
2045                 static $table = [
2046                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2047                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2048                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2049                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2050                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2051                 ];
2052
2053                 $num = intval( $num );
2054                 if ( $num > 10000 || $num <= 0 ) {
2055                         return $num;
2056                 }
2057
2058                 $s = '';
2059                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2060                         if ( $num >= $pow10 ) {
2061                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2062                         }
2063                         $num = $num % $pow10;
2064                 }
2065                 return $s;
2066         }
2067
2068         /**
2069          * Hebrew Gematria number formatting up to 9999
2070          *
2071          * @param int $num
2072          *
2073          * @return string
2074          */
2075         static function hebrewNumeral( $num ) {
2076                 static $table = [
2077                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2078                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2079                         [ '',
2080                                 [ 'ק' ],
2081                                 [ 'ר' ],
2082                                 [ 'ש' ],
2083                                 [ 'ת' ],
2084                                 [ 'ת', 'ק' ],
2085                                 [ 'ת', 'ר' ],
2086                                 [ 'ת', 'ש' ],
2087                                 [ 'ת', 'ת' ],
2088                                 [ 'ת', 'ת', 'ק' ],
2089                                 [ 'ת', 'ת', 'ר' ],
2090                         ],
2091                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2092                 ];
2093
2094                 $num = intval( $num );
2095                 if ( $num > 9999 || $num <= 0 ) {
2096                         return $num;
2097                 }
2098
2099                 // Round thousands have special notations
2100                 if ( $num === 1000 ) {
2101                         return "א' אלף";
2102                 } elseif ( $num % 1000 === 0 ) {
2103                         return $table[0][$num / 1000] . "' אלפים";
2104                 }
2105
2106                 $letters = [];
2107
2108                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2109                         if ( $num >= $pow10 ) {
2110                                 if ( $num === 15 || $num === 16 ) {
2111                                         $letters[] = $table[0][9];
2112                                         $letters[] = $table[0][$num - 9];
2113                                         $num = 0;
2114                                 } else {
2115                                         $letters = array_merge(
2116                                                 $letters,
2117                                                 (array)$table[$i][intval( $num / $pow10 )]
2118                                         );
2119
2120                                         if ( $pow10 === 1000 ) {
2121                                                 $letters[] = "'";
2122                                         }
2123                                 }
2124                         }
2125
2126                         $num = $num % $pow10;
2127                 }
2128
2129                 $preTransformLength = count( $letters );
2130                 if ( $preTransformLength === 1 ) {
2131                         // Add geresh (single quote) to one-letter numbers
2132                         $letters[] = "'";
2133                 } else {
2134                         $lastIndex = $preTransformLength - 1;
2135                         $letters[$lastIndex] = str_replace(
2136                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2137                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2138                                 $letters[$lastIndex]
2139                         );
2140
2141                         // Add gershayim (double quote) to multiple-letter numbers,
2142                         // but exclude numbers with only one letter after the thousands
2143                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2144                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2145                                 $letters[] = "'";
2146                         } else {
2147                                 array_splice( $letters, -1, 0, '"' );
2148                         }
2149                 }
2150
2151                 return implode( $letters );
2152         }
2153
2154         /**
2155          * Used by date() and time() to adjust the time output.
2156          *
2157          * @param string $ts The time in date('YmdHis') format
2158          * @param mixed $tz Adjust the time by this amount (default false, mean we
2159          *   get user timecorrection setting)
2160          * @return int
2161          */
2162         public function userAdjust( $ts, $tz = false ) {
2163                 global $wgUser, $wgLocalTZoffset;
2164
2165                 if ( $tz === false ) {
2166                         $tz = $wgUser->getOption( 'timecorrection' );
2167                 }
2168
2169                 $data = explode( '|', $tz, 3 );
2170
2171                 if ( $data[0] == 'ZoneInfo' ) {
2172                         try {
2173                                 $userTZ = new DateTimeZone( $data[2] );
2174                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2175                                 $date->setTimezone( $userTZ );
2176                                 return $date->format( 'YmdHis' );
2177                         } catch ( Exception $e ) {
2178                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2179                                 $data[0] = 'Offset';
2180                         }
2181                 }
2182
2183                 if ( $data[0] == 'System' || $tz == '' ) {
2184                         # Global offset in minutes.
2185                         $minDiff = $wgLocalTZoffset;
2186                 } elseif ( $data[0] == 'Offset' ) {
2187                         $minDiff = intval( $data[1] );
2188                 } else {
2189                         $data = explode( ':', $tz );
2190                         if ( count( $data ) == 2 ) {
2191                                 $data[0] = intval( $data[0] );
2192                                 $data[1] = intval( $data[1] );
2193                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2194                                 if ( $data[0] < 0 ) {
2195                                         $minDiff = -$minDiff;
2196                                 }
2197                         } else {
2198                                 $minDiff = intval( $data[0] ) * 60;
2199                         }
2200                 }
2201
2202                 # No difference ? Return time unchanged
2203                 if ( 0 == $minDiff ) {
2204                         return $ts;
2205                 }
2206
2207                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2208                 # Generate an adjusted date; take advantage of the fact that mktime
2209                 # will normalize out-of-range values so we don't have to split $minDiff
2210                 # into hours and minutes.
2211                 $t = mktime( (
2212                         (int)substr( $ts, 8, 2 ) ), # Hours
2213                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2214                         (int)substr( $ts, 12, 2 ), # Seconds
2215                         (int)substr( $ts, 4, 2 ), # Month
2216                         (int)substr( $ts, 6, 2 ), # Day
2217                         (int)substr( $ts, 0, 4 ) ); # Year
2218
2219                 $date = date( 'YmdHis', $t );
2220                 Wikimedia\restoreWarnings();
2221
2222                 return $date;
2223         }
2224
2225         /**
2226          * This is meant to be used by time(), date(), and timeanddate() to get
2227          * the date preference they're supposed to use, it should be used in
2228          * all children.
2229          *
2230          *     function timeanddate([...], $format = true) {
2231          *       $datePreference = $this->dateFormat($format);
2232          *       [...]
2233          *     }
2234          *
2235          * @param int|string|bool $usePrefs If true, the user's preference is used
2236          *   if false, the site/language default is used
2237          *   if int/string, assumed to be a format.
2238          * @return string
2239          */
2240         function dateFormat( $usePrefs = true ) {
2241                 global $wgUser;
2242
2243                 if ( is_bool( $usePrefs ) ) {
2244                         if ( $usePrefs ) {
2245                                 $datePreference = $wgUser->getDatePreference();
2246                         } else {
2247                                 $datePreference = (string)User::getDefaultOption( 'date' );
2248                         }
2249                 } else {
2250                         $datePreference = (string)$usePrefs;
2251                 }
2252
2253                 // return int
2254                 if ( $datePreference == '' ) {
2255                         return 'default';
2256                 }
2257
2258                 return $datePreference;
2259         }
2260
2261         /**
2262          * Get a format string for a given type and preference
2263          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2264          * @param string $pref The format name as it appears in Messages*.php under
2265          *  $datePreferences.
2266          *
2267          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2268          *
2269          * @return string
2270          */
2271         function getDateFormatString( $type, $pref ) {
2272                 $wasDefault = false;
2273                 if ( $pref == 'default' ) {
2274                         $wasDefault = true;
2275                         $pref = $this->getDefaultDateFormat();
2276                 }
2277
2278                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2279                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2280
2281                         if ( $type === 'pretty' && $df === null ) {
2282                                 $df = $this->getDateFormatString( 'date', $pref );
2283                         }
2284
2285                         if ( !$wasDefault && $df === null ) {
2286                                 $pref = $this->getDefaultDateFormat();
2287                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2288                         }
2289
2290                         $this->dateFormatStrings[$type][$pref] = $df;
2291                 }
2292                 return $this->dateFormatStrings[$type][$pref];
2293         }
2294
2295         /**
2296          * @param string $ts The time format which needs to be turned into a
2297          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2298          * @param bool $adj Whether to adjust the time output according to the
2299          *   user configured offset ($timecorrection)
2300          * @param mixed $format True to use user's date format preference
2301          * @param string|bool $timecorrection The time offset as returned by
2302          *   validateTimeZone() in Special:Preferences
2303          * @return string
2304          */
2305         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2306                 $ts = wfTimestamp( TS_MW, $ts );
2307                 if ( $adj ) {
2308                         $ts = $this->userAdjust( $ts, $timecorrection );
2309                 }
2310                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2311                 return $this->sprintfDate( $df, $ts );
2312         }
2313
2314         /**
2315          * @param string $ts The time format which needs to be turned into a
2316          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2317          * @param bool $adj Whether to adjust the time output according to the
2318          *   user configured offset ($timecorrection)
2319          * @param mixed $format True to use user's date format preference
2320          * @param string|bool $timecorrection The time offset as returned by
2321          *   validateTimeZone() in Special:Preferences
2322          * @return string
2323          */
2324         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2325                 $ts = wfTimestamp( TS_MW, $ts );
2326                 if ( $adj ) {
2327                         $ts = $this->userAdjust( $ts, $timecorrection );
2328                 }
2329                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2330                 return $this->sprintfDate( $df, $ts );
2331         }
2332
2333         /**
2334          * @param string $ts The time format which needs to be turned into a
2335          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2336          * @param bool $adj Whether to adjust the time output according to the
2337          *   user configured offset ($timecorrection)
2338          * @param mixed $format What format to return, if it's false output the
2339          *   default one (default true)
2340          * @param string|bool $timecorrection The time offset as returned by
2341          *   validateTimeZone() in Special:Preferences
2342          * @return string
2343          */
2344         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2345                 $ts = wfTimestamp( TS_MW, $ts );
2346                 if ( $adj ) {
2347                         $ts = $this->userAdjust( $ts, $timecorrection );
2348                 }
2349                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2350                 return $this->sprintfDate( $df, $ts );
2351         }
2352
2353         /**
2354          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2355          *
2356          * @since 1.20
2357          *
2358          * @param int $seconds The amount of seconds.
2359          * @param array $chosenIntervals The intervals to enable.
2360          *
2361          * @return string
2362          */
2363         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2364                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2365
2366                 $segments = [];
2367
2368                 foreach ( $intervals as $intervalName => $intervalValue ) {
2369                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2370                         // duration-years, duration-decades, duration-centuries, duration-millennia
2371                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2372                         $segments[] = $message->inLanguage( $this )->escaped();
2373                 }
2374
2375                 return $this->listToText( $segments );
2376         }
2377
2378         /**
2379          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2380          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2381          *
2382          * @since 1.20
2383          *
2384          * @param int $seconds The amount of seconds.
2385          * @param array $chosenIntervals The intervals to enable.
2386          *
2387          * @return array
2388          */
2389         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2390                 if ( empty( $chosenIntervals ) ) {
2391                         $chosenIntervals = [
2392                                 'millennia',
2393                                 'centuries',
2394                                 'decades',
2395                                 'years',
2396                                 'days',
2397                                 'hours',
2398                                 'minutes',
2399                                 'seconds'
2400                         ];
2401                 }
2402
2403                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2404                 $sortedNames = array_keys( $intervals );
2405                 $smallestInterval = array_pop( $sortedNames );
2406
2407                 $segments = [];
2408
2409                 foreach ( $intervals as $name => $length ) {
2410                         $value = floor( $seconds / $length );
2411
2412                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2413                                 $seconds -= $value * $length;
2414                                 $segments[$name] = $value;
2415                         }
2416                 }
2417
2418                 return $segments;
2419         }
2420
2421         /**
2422          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2423          *
2424          * @param string $type Can be 'date', 'time' or 'both'
2425          * @param string $ts The time format which needs to be turned into a
2426          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2427          * @param User $user User object used to get preferences for timezone and format
2428          * @param array $options Array, can contain the following keys:
2429          *   - 'timecorrection': time correction, can have the following values:
2430          *     - true: use user's preference
2431          *     - false: don't use time correction
2432          *     - int: value of time correction in minutes
2433          *   - 'format': format to use, can have the following values:
2434          *     - true: use user's preference
2435          *     - false: use default preference
2436          *     - string: format to use
2437          * @since 1.19
2438          * @return string
2439          */
2440         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2441                 $ts = wfTimestamp( TS_MW, $ts );
2442                 $options += [ 'timecorrection' => true, 'format' => true ];
2443                 if ( $options['timecorrection'] !== false ) {
2444                         if ( $options['timecorrection'] === true ) {
2445                                 $offset = $user->getOption( 'timecorrection' );
2446                         } else {
2447                                 $offset = $options['timecorrection'];
2448                         }
2449                         $ts = $this->userAdjust( $ts, $offset );
2450                 }
2451                 if ( $options['format'] === true ) {
2452                         $format = $user->getDatePreference();
2453                 } else {
2454                         $format = $options['format'];
2455                 }
2456                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2457                 return $this->sprintfDate( $df, $ts );
2458         }
2459
2460         /**
2461          * Get the formatted date for the given timestamp and formatted for
2462          * the given user.
2463          *
2464          * @param mixed $ts Mixed: the time format which needs to be turned into a
2465          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2466          * @param User $user User object used to get preferences for timezone and format
2467          * @param array $options Array, can contain the following keys:
2468          *   - 'timecorrection': time correction, can have the following values:
2469          *     - true: use user's preference
2470          *     - false: don't use time correction
2471          *     - int: value of time correction in minutes
2472          *   - 'format': format to use, can have the following values:
2473          *     - true: use user's preference
2474          *     - false: use default preference
2475          *     - string: format to use
2476          * @since 1.19
2477          * @return string
2478          */
2479         public function userDate( $ts, User $user, array $options = [] ) {
2480                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2481         }
2482
2483         /**
2484          * Get the formatted time for the given timestamp and formatted for
2485          * the given user.
2486          *
2487          * @param mixed $ts The time format which needs to be turned into a
2488          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2489          * @param User $user User object used to get preferences for timezone and format
2490          * @param array $options Array, can contain the following keys:
2491          *   - 'timecorrection': time correction, can have the following values:
2492          *     - true: use user's preference
2493          *     - false: don't use time correction
2494          *     - int: value of time correction in minutes
2495          *   - 'format': format to use, can have the following values:
2496          *     - true: use user's preference
2497          *     - false: use default preference
2498          *     - string: format to use
2499          * @since 1.19
2500          * @return string
2501          */
2502         public function userTime( $ts, User $user, array $options = [] ) {
2503                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2504         }
2505
2506         /**
2507          * Get the formatted date and time for the given timestamp and formatted for
2508          * the given user.
2509          *
2510          * @param mixed $ts The time format which needs to be turned into a
2511          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2512          * @param User $user User object used to get preferences for timezone and format
2513          * @param array $options Array, can contain the following keys:
2514          *   - 'timecorrection': time correction, can have the following values:
2515          *     - true: use user's preference
2516          *     - false: don't use time correction
2517          *     - int: value of time correction in minutes
2518          *   - 'format': format to use, can have the following values:
2519          *     - true: use user's preference
2520          *     - false: use default preference
2521          *     - string: format to use
2522          * @since 1.19
2523          * @return string
2524          */
2525         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2526                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2527         }
2528
2529         /**
2530          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2531          *
2532          * Determine the difference between the timestamp and the current time, and
2533          * generate a readable timestamp by returning "<N> <units> ago", where the
2534          * largest possible unit is used.
2535          *
2536          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2537          *
2538          * @param MWTimestamp $time
2539          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2540          * @param User|null $user User the timestamp is being generated for
2541          *  (or null to use main context's user)
2542          * @return string Formatted timestamp
2543          */
2544         public function getHumanTimestamp(
2545                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2546         ) {
2547                 if ( $relativeTo === null ) {
2548                         $relativeTo = new MWTimestamp();
2549                 }
2550                 if ( $user === null ) {
2551                         $user = RequestContext::getMain()->getUser();
2552                 }
2553
2554                 // Adjust for the user's timezone.
2555                 $offsetThis = $time->offsetForUser( $user );
2556                 $offsetRel = $relativeTo->offsetForUser( $user );
2557
2558                 $ts = '';
2559                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2560                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2561                 }
2562
2563                 // Reset the timezone on the objects.
2564                 $time->timestamp->sub( $offsetThis );
2565                 $relativeTo->timestamp->sub( $offsetRel );
2566
2567                 return $ts;
2568         }
2569
2570         /**
2571          * Convert an MWTimestamp into a pretty human-readable timestamp using
2572          * the given user preferences and relative base time.
2573          *
2574          * @see Language::getHumanTimestamp
2575          * @param MWTimestamp $ts Timestamp to prettify
2576          * @param MWTimestamp $relativeTo Base timestamp
2577          * @param User $user User preferences to use
2578          * @return string Human timestamp
2579          * @since 1.26
2580          */
2581         private function getHumanTimestampInternal(
2582                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2583         ) {
2584                 $diff = $ts->diff( $relativeTo );
2585                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2586                         (int)$relativeTo->timestamp->format( 'w' ) );
2587                 $days = $diff->days ?: (int)$diffDay;
2588                 if ( $diff->invert || $days > 5
2589                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2590                 ) {
2591                         // Timestamps are in different years: use full timestamp
2592                         // Also do full timestamp for future dates
2593                         /**
2594                          * @todo FIXME: Add better handling of future timestamps.
2595                          */
2596                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2597                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2598                 } elseif ( $days > 5 ) {
2599                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2600                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2601                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2602                 } elseif ( $days > 1 ) {
2603                         // Timestamp within the past week: show the day of the week and time
2604                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2605                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2606                         // Messages:
2607                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2608                         $ts = wfMessage( "$weekday-at" )
2609                                 ->inLanguage( $this )
2610                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2611                                 ->text();
2612                 } elseif ( $days == 1 ) {
2613                         // Timestamp was yesterday: say 'yesterday' and the time.
2614                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2615                         $ts = wfMessage( 'yesterday-at' )
2616                                 ->inLanguage( $this )
2617                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2618                                 ->text();
2619                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2620                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2621                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2622                         $ts = wfMessage( 'today-at' )
2623                                 ->inLanguage( $this )
2624                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2625                                 ->text();
2626
2627                 // From here on in, the timestamp was soon enough ago so that we can simply say
2628                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2629                 } elseif ( $diff->h == 1 ) {
2630                         // Less than 90 minutes, but more than an hour ago.
2631                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2632                 } elseif ( $diff->i >= 1 ) {
2633                         // A few minutes ago.
2634                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2635                 } elseif ( $diff->s >= 30 ) {
2636                         // Less than a minute, but more than 30 sec ago.
2637                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2638                 } else {
2639                         // Less than 30 seconds ago.
2640                         $ts = wfMessage( 'just-now' )->text();
2641                 }
2642
2643                 return $ts;
2644         }
2645
2646         /**
2647          * @param string $key
2648          * @return string|null
2649          */
2650         public function getMessage( $key ) {
2651                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2652         }
2653
2654         /**
2655          * @return array
2656          */
2657         function getAllMessages() {
2658                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2659         }
2660
2661         /**
2662          * @param string $in
2663          * @param string $out
2664          * @param string $string
2665          * @return string
2666          */
2667         public function iconv( $in, $out, $string ) {
2668                 # Even with //IGNORE iconv can whine about illegal characters in
2669                 # *input* string. We just ignore those too.
2670                 # REF: https://bugs.php.net/bug.php?id=37166
2671                 # REF: https://phabricator.wikimedia.org/T18885
2672                 Wikimedia\suppressWarnings();
2673                 $text = iconv( $in, $out . '//IGNORE', $string );
2674                 Wikimedia\restoreWarnings();
2675                 return $text;
2676         }
2677
2678         // callback functions for ucwords(), ucwordbreaks()
2679
2680         /**
2681          * @param array $matches
2682          * @return mixed|string
2683          */
2684         function ucwordbreaksCallbackAscii( $matches ) {
2685                 return $this->ucfirst( $matches[1] );
2686         }
2687
2688         /**
2689          * @param array $matches
2690          * @return string
2691          */
2692         function ucwordbreaksCallbackMB( $matches ) {
2693                 return mb_strtoupper( $matches[0] );
2694         }
2695
2696         /**
2697          * @param array $matches
2698          * @return string
2699          */
2700         function ucwordsCallbackMB( $matches ) {
2701                 return mb_strtoupper( $matches[0] );
2702         }
2703
2704         /**
2705          * Make a string's first character uppercase
2706          *
2707          * @param string $str
2708          *
2709          * @return string
2710          */
2711         public function ucfirst( $str ) {
2712                 $o = ord( $str );
2713                 if ( $o < 96 ) { // if already uppercase...
2714                         return $str;
2715                 } elseif ( $o < 128 ) {
2716                         return ucfirst( $str ); // use PHP's ucfirst()
2717                 } else {
2718                         // fall back to more complex logic in case of multibyte strings
2719                         return $this->uc( $str, true );
2720                 }
2721         }
2722
2723         /**
2724          * Convert a string to uppercase
2725          *
2726          * @param string $str
2727          * @param bool $first
2728          *
2729          * @return string
2730          */
2731         public function uc( $str, $first = false ) {
2732                 if ( $first ) {
2733                         if ( $this->isMultibyte( $str ) ) {
2734                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2735                         } else {
2736                                 return ucfirst( $str );
2737                         }
2738                 } else {
2739                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2740                 }
2741         }
2742
2743         /**
2744          * @param string $str
2745          * @return mixed|string
2746          */
2747         function lcfirst( $str ) {
2748                 $o = ord( $str );
2749                 if ( !$o ) {
2750                         return strval( $str );
2751                 } elseif ( $o >= 128 ) {
2752                         return $this->lc( $str, true );
2753                 } elseif ( $o > 96 ) {
2754                         return $str;
2755                 } else {
2756                         $str[0] = strtolower( $str[0] );
2757                         return $str;
2758                 }
2759         }
2760
2761         /**
2762          * @param string $str
2763          * @param bool $first
2764          * @return mixed|string
2765          */
2766         function lc( $str, $first = false ) {
2767                 if ( $first ) {
2768                         if ( $this->isMultibyte( $str ) ) {
2769                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2770                         } else {
2771                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2772                         }
2773                 } else {
2774                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2775                 }
2776         }
2777
2778         /**
2779          * @param string $str
2780          * @return bool
2781          */
2782         function isMultibyte( $str ) {
2783                 return strlen( $str ) !== mb_strlen( $str );
2784         }
2785
2786         /**
2787          * @param string $str
2788          * @return mixed|string
2789          */
2790         function ucwords( $str ) {
2791                 if ( $this->isMultibyte( $str ) ) {
2792                         $str = $this->lc( $str );
2793
2794                         // regexp to find first letter in each word (i.e. after each space)
2795                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2796
2797                         // function to use to capitalize a single char
2798                         return preg_replace_callback(
2799                                 $replaceRegexp,
2800                                 [ $this, 'ucwordsCallbackMB' ],
2801                                 $str
2802                         );
2803                 } else {
2804                         return ucwords( strtolower( $str ) );
2805                 }
2806         }
2807
2808         /**
2809          * capitalize words at word breaks
2810          *
2811          * @param string $str
2812          * @return mixed
2813          */
2814         function ucwordbreaks( $str ) {
2815                 if ( $this->isMultibyte( $str ) ) {
2816                         $str = $this->lc( $str );
2817
2818                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2819                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2820
2821                         // find first letter after word break
2822                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2823                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2824
2825                         return preg_replace_callback(
2826                                 $replaceRegexp,
2827                                 [ $this, 'ucwordbreaksCallbackMB' ],
2828                                 $str
2829                         );
2830                 } else {
2831                         return preg_replace_callback(
2832                                 '/\b([\w\x80-\xff]+)\b/',
2833                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2834                                 $str
2835                         );
2836                 }
2837         }
2838
2839         /**
2840          * Return a case-folded representation of $s
2841          *
2842          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2843          * and $s2 are the same except for the case of their characters. It is not
2844          * necessary for the value returned to make sense when displayed.
2845          *
2846          * Do *not* perform any other normalisation in this function. If a caller
2847          * uses this function when it should be using a more general normalisation
2848          * function, then fix the caller.
2849          *
2850          * @param string $s
2851          *
2852          * @return string
2853          */
2854         function caseFold( $s ) {
2855                 return $this->uc( $s );
2856         }
2857
2858         /**
2859          * @param string $s
2860          * @return string
2861          * @throws MWException
2862          */
2863         function checkTitleEncoding( $s ) {
2864                 if ( is_array( $s ) ) {
2865                         throw new MWException( 'Given array to checkTitleEncoding.' );
2866                 }
2867                 if ( StringUtils::isUtf8( $s ) ) {
2868                         return $s;
2869                 }
2870
2871                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2872         }
2873
2874         /**
2875          * @return array
2876          */
2877         function fallback8bitEncoding() {
2878                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2879         }
2880
2881         /**
2882          * Most writing systems use whitespace to break up words.
2883          * Some languages such as Chinese don't conventionally do this,
2884          * which requires special handling when breaking up words for
2885          * searching etc.
2886          *
2887          * @return bool
2888          */
2889         function hasWordBreaks() {
2890                 return true;
2891         }
2892
2893         /**
2894          * Some languages such as Chinese require word segmentation,
2895          * Specify such segmentation when overridden in derived class.
2896          *
2897          * @param string $string
2898          * @return string
2899          */
2900         function segmentByWord( $string ) {
2901                 return $string;
2902         }
2903
2904         /**
2905          * Some languages have special punctuation need to be normalized.
2906          * Make such changes here.
2907          *
2908          * @param string $string
2909          * @return string
2910          */
2911         function normalizeForSearch( $string ) {
2912                 return self::convertDoubleWidth( $string );
2913         }
2914
2915         /**
2916          * convert double-width roman characters to single-width.
2917          * range: ff00-ff5f ~= 0020-007f
2918          *
2919          * @param string $string
2920          *
2921          * @return string
2922          */
2923         protected static function convertDoubleWidth( $string ) {
2924                 static $full = null;
2925                 static $half = null;
2926
2927                 if ( $full === null ) {
2928                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2929                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2930                         $full = str_split( $fullWidth, 3 );
2931                         $half = str_split( $halfWidth );
2932                 }
2933
2934                 $string = str_replace( $full, $half, $string );
2935                 return $string;
2936         }
2937
2938         /**
2939          * @param string $string
2940          * @param string $pattern
2941          * @return string
2942          */
2943         protected static function insertSpace( $string, $pattern ) {
2944                 $string = preg_replace( $pattern, " $1 ", $string );
2945                 $string = preg_replace( '/ +/', ' ', $string );
2946                 return $string;
2947         }
2948
2949         /**
2950          * @param array $termsArray
2951          * @return array
2952          */
2953         function convertForSearchResult( $termsArray ) {
2954                 # some languages, e.g. Chinese, need to do a conversion
2955                 # in order for search results to be displayed correctly
2956                 return $termsArray;
2957         }
2958
2959         /**
2960          * Get the first character of a string.
2961          *
2962          * @param string $s
2963          * @return string
2964          */
2965         function firstChar( $s ) {
2966                 $matches = [];
2967                 preg_match(
2968                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2969                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2970                         $s,
2971                         $matches
2972                 );
2973
2974                 if ( isset( $matches[1] ) ) {
2975                         if ( strlen( $matches[1] ) != 3 ) {
2976                                 return $matches[1];
2977                         }
2978
2979                         // Break down Hangul syllables to grab the first jamo
2980                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2981                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2982                                 return $matches[1];
2983                         } elseif ( $code < 0xb098 ) {
2984                                 return "\u{3131}";
2985                         } elseif ( $code < 0xb2e4 ) {
2986                                 return "\u{3134}";
2987                         } elseif ( $code < 0xb77c ) {
2988                                 return "\u{3137}";
2989                         } elseif ( $code < 0xb9c8 ) {
2990                                 return "\u{3139}";
2991                         } elseif ( $code < 0xbc14 ) {
2992                                 return "\u{3141}";
2993                         } elseif ( $code < 0xc0ac ) {
2994                                 return "\u{3142}";
2995                         } elseif ( $code < 0xc544 ) {
2996                                 return "\u{3145}";
2997                         } elseif ( $code < 0xc790 ) {
2998                                 return "\u{3147}";
2999                         } elseif ( $code < 0xcc28 ) {
3000                                 return "\u{3148}";
3001                         } elseif ( $code < 0xce74 ) {
3002                                 return "\u{314A}";
3003                         } elseif ( $code < 0xd0c0 ) {
3004                                 return "\u{314B}";
3005                         } elseif ( $code < 0xd30c ) {
3006                                 return "\u{314C}";
3007                         } elseif ( $code < 0xd558 ) {
3008                                 return "\u{314D}";
3009                         } else {
3010                                 return "\u{314E}";
3011                         }
3012                 } else {
3013                         return '';
3014                 }
3015         }
3016
3017         /**
3018          * @deprecated No-op since 1.28
3019          */
3020         function initEncoding() {
3021                 wfDeprecated( __METHOD__, '1.28' );
3022                 // No-op.
3023         }
3024
3025         /**
3026          * @param string $s
3027          * @return string
3028          * @deprecated No-op since 1.28
3029          */
3030         function recodeForEdit( $s ) {
3031                 wfDeprecated( __METHOD__, '1.28' );
3032                 return $s;
3033         }
3034
3035         /**
3036          * @param string $s
3037          * @return string
3038          * @deprecated No-op since 1.28
3039          */
3040         function recodeInput( $s ) {
3041                 wfDeprecated( __METHOD__, '1.28' );
3042                 return $s;
3043         }
3044
3045         /**
3046          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3047          * also cleans up certain backwards-compatible sequences, converting them
3048          * to the modern Unicode equivalent.
3049          *
3050          * This is language-specific for performance reasons only.
3051          *
3052          * @param string $s
3053          *
3054          * @return string
3055          */
3056         public function normalize( $s ) {
3057                 global $wgAllUnicodeFixes;
3058                 $s = UtfNormal\Validator::cleanUp( $s );
3059                 if ( $wgAllUnicodeFixes ) {
3060                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3061                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3062                 }
3063
3064                 return $s;
3065         }
3066
3067         /**
3068          * Transform a string using serialized data stored in the given file (which
3069          * must be in the serialized subdirectory of $IP). The file contains pairs
3070          * mapping source characters to destination characters.
3071          *
3072          * The data is cached in process memory. This will go faster if you have the
3073          * FastStringSearch extension.
3074          *
3075          * @param string $file
3076          * @param string $string
3077          *
3078          * @throws MWException
3079          * @return string
3080          */
3081         protected function transformUsingPairFile( $file, $string ) {
3082                 if ( !isset( $this->transformData[$file] ) ) {
3083                         global $IP;
3084                         $data = require "$IP/languages/data/{$file}";
3085                         $this->transformData[$file] = new ReplacementArray( $data );
3086                 }
3087                 return $this->transformData[$file]->replace( $string );
3088         }
3089
3090         /**
3091          * For right-to-left language support
3092          *
3093          * @return bool
3094          */
3095         function isRTL() {
3096                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3097         }
3098
3099         /**
3100          * Return the correct HTML 'dir' attribute value for this language.
3101          * @return string
3102          */
3103         function getDir() {
3104                 return $this->isRTL() ? 'rtl' : 'ltr';
3105         }
3106
3107         /**
3108          * Return 'left' or 'right' as appropriate alignment for line-start
3109          * for this language's text direction.
3110          *
3111          * Should be equivalent to CSS3 'start' text-align value....
3112          *
3113          * @return string
3114          */
3115         function alignStart() {
3116                 return $this->isRTL() ? 'right' : 'left';
3117         }
3118
3119         /**
3120          * Return 'right' or 'left' as appropriate alignment for line-end
3121          * for this language's text direction.
3122          *
3123          * Should be equivalent to CSS3 'end' text-align value....
3124          *
3125          * @return string
3126          */
3127         function alignEnd() {
3128                 return $this->isRTL() ? 'left' : 'right';
3129         }
3130
3131         /**
3132          * A hidden direction mark (LRM or RLM), depending on the language direction.
3133          * Unlike getDirMark(), this function returns the character as an HTML entity.
3134          * This function should be used when the output is guaranteed to be HTML,
3135          * because it makes the output HTML source code more readable. When
3136          * the output is plain text or can be escaped, getDirMark() should be used.
3137          *
3138          * @param bool $opposite Get the direction mark opposite to your language
3139          * @return string
3140          * @since 1.20
3141          */
3142         function getDirMarkEntity( $opposite = false ) {
3143                 if ( $opposite ) {
3144                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3145                 }
3146                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3147         }
3148
3149         /**
3150          * A hidden direction mark (LRM or RLM), depending on the language direction.
3151          * This function produces them as invisible Unicode characters and
3152          * the output may be hard to read and debug, so it should only be used
3153          * when the output is plain text or can be escaped. When the output is
3154          * HTML, use getDirMarkEntity() instead.
3155          *
3156          * @param bool $opposite Get the direction mark opposite to your language
3157          * @return string
3158          */
3159         function getDirMark( $opposite = false ) {
3160                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3161                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3162                 if ( $opposite ) {
3163                         return $this->isRTL() ? $lrm : $rlm;
3164                 }
3165                 return $this->isRTL() ? $rlm : $lrm;
3166         }
3167
3168         /**
3169          * @return array
3170          */
3171         function capitalizeAllNouns() {
3172                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3173         }
3174
3175         /**
3176          * An arrow, depending on the language direction.
3177          *
3178          * @param string $direction The direction of the arrow: forwards (default),
3179          *   backwards, left, right, up, down.
3180          * @return string
3181          */
3182         function getArrow( $direction = 'forwards' ) {
3183                 switch ( $direction ) {
3184                         case 'forwards':
3185                                 return $this->isRTL() ? '←' : '→';
3186                         case 'backwards':
3187                                 return $this->isRTL() ? '→' : '←';
3188                         case 'left':
3189                                 return '←';
3190                         case 'right':
3191                                 return '→';
3192                         case 'up':
3193                                 return '↑';
3194                         case 'down':
3195                                 return '↓';
3196                 }
3197         }
3198
3199         /**
3200          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3201          *
3202          * @return bool
3203          */
3204         function linkPrefixExtension() {
3205                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3206         }
3207
3208         /**
3209          * Get all magic words from cache.
3210          * @return array
3211          */
3212         function getMagicWords() {
3213                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3214         }
3215
3216         /**
3217          * Run the LanguageGetMagic hook once.
3218          */
3219         protected function doMagicHook() {
3220                 if ( $this->mMagicHookDone ) {
3221                         return;
3222                 }
3223                 $this->mMagicHookDone = true;
3224                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ], '1.16' );
3225         }
3226
3227         /**
3228          * Fill a MagicWord object with data from here
3229          *
3230          * @param MagicWord $mw
3231          */
3232         function getMagic( $mw ) {
3233                 // Saves a function call
3234                 if ( !$this->mMagicHookDone ) {
3235                         $this->doMagicHook();
3236                 }
3237
3238                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3239                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3240                 } else {
3241                         $rawEntry = self::$dataCache->getSubitem(
3242                                 $this->mCode, 'magicWords', $mw->mId );
3243                 }
3244
3245                 if ( !is_array( $rawEntry ) ) {
3246                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3247                 } else {
3248                         $mw->mCaseSensitive = $rawEntry[0];
3249                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3250                 }
3251         }
3252
3253         /**
3254          * Add magic words to the extension array
3255          *
3256          * @param array $newWords
3257          */
3258         function addMagicWordsByLang( $newWords ) {
3259                 $fallbackChain = $this->getFallbackLanguages();
3260                 $fallbackChain = array_reverse( $fallbackChain );
3261                 foreach ( $fallbackChain as $code ) {
3262                         if ( isset( $newWords[$code] ) ) {
3263                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3264                         }
3265                 }
3266         }
3267
3268         /**
3269          * Get special page names, as an associative array
3270          *   canonical name => array of valid names, including aliases
3271          * @return array
3272          */
3273         function getSpecialPageAliases() {
3274                 // Cache aliases because it may be slow to load them
3275                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3276                         // Initialise array
3277                         $this->mExtendedSpecialPageAliases =
3278                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3279                         Hooks::run( 'LanguageGetSpecialPageAliases',
3280                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ], '1.16' );
3281                 }
3282
3283                 return $this->mExtendedSpecialPageAliases;
3284         }
3285
3286         /**
3287          * Italic is unsuitable for some languages
3288          *
3289          * @param string $text The text to be emphasized.
3290          * @return string
3291          */
3292         function emphasize( $text ) {
3293                 return "<em>$text</em>";
3294         }
3295
3296         /**
3297          * Normally we output all numbers in plain en_US style, that is
3298          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3299          * point twohundredthirtyfive. However this is not suitable for all
3300          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3301          * Icelandic just want to use commas instead of dots, and dots instead
3302          * of commas like "293.291,235".
3303          *
3304          * An example of this function being called:
3305          * <code>
3306          * wfMessage( 'message' )->numParams( $num )->text()
3307          * </code>
3308          *
3309          * See $separatorTransformTable on MessageIs.php for
3310          * the , => . and . => , implementation.
3311          *
3312          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3313          * @param int|float $number The string to be formatted, should be an integer
3314          *   or a floating point number.
3315          * @param bool $nocommafy Set to true for special numbers like dates
3316          * @return string
3317          */
3318         public function formatNum( $number, $nocommafy = false ) {
3319                 global $wgTranslateNumerals;
3320                 if ( !$nocommafy ) {
3321                         $number = $this->commafy( $number );
3322                         $s = $this->separatorTransformTable();
3323                         if ( $s ) {
3324                                 $number = strtr( $number, $s );
3325                         }
3326                 }
3327
3328                 if ( $wgTranslateNumerals ) {
3329                         $s = $this->digitTransformTable();
3330                         if ( $s ) {
3331                                 $number = strtr( $number, $s );
3332                         }
3333                 }
3334
3335                 return (string)$number;
3336         }
3337
3338         /**
3339          * Front-end for non-commafied formatNum
3340          *
3341          * @param int|float $number The string to be formatted, should be an integer
3342          *        or a floating point number.
3343          * @since 1.21
3344          * @return string
3345          */
3346         public function formatNumNoSeparators( $number ) {
3347                 return $this->formatNum( $number, true );
3348         }
3349
3350         /**
3351          * @param string $number
3352          * @return string
3353          */
3354         public function parseFormattedNumber( $number ) {
3355                 $s = $this->digitTransformTable();
3356                 if ( $s ) {
3357                         // eliminate empty array values such as ''. (T66347)
3358                         $s = array_filter( $s );
3359                         $number = strtr( $number, array_flip( $s ) );
3360                 }
3361
3362                 $s = $this->separatorTransformTable();
3363                 if ( $s ) {
3364                         // eliminate empty array values such as ''. (T66347)
3365                         $s = array_filter( $s );
3366                         $number = strtr( $number, array_flip( $s ) );
3367                 }
3368
3369                 $number = strtr( $number, [ ',' => '' ] );
3370                 return $number;
3371         }
3372
3373         /**
3374          * Adds commas to a given number
3375          * @since 1.19
3376          * @param mixed $number
3377          * @return string
3378          */
3379         function commafy( $number ) {
3380                 $digitGroupingPattern = $this->digitGroupingPattern();
3381                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3382                 if ( $number === null ) {
3383                         return '';
3384                 }
3385
3386                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3387                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3388                         // In some languages it's conventional not to insert a thousands separator
3389                         // in numbers that are four digits long (1000-9999).
3390                         if ( $minimumGroupingDigits ) {
3391                                 // Number of '#' characters after last comma in the grouping pattern.
3392                                 // The pattern is hardcoded here, but this would vary for different patterns.
3393                                 $primaryGroupingSize = 3;
3394                                 // Maximum length of a number to suppress digit grouping for.
3395                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3396                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3397                                         return $number;
3398                                 }
3399                         }
3400                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3401                 } else {
3402                         // Ref: http://cldr.unicode.org/translation/number-patterns
3403                         $sign = "";
3404                         if ( intval( $number ) < 0 ) {
3405                                 // For negative numbers apply the algorithm like positive number and add sign.
3406                                 $sign = "-";
3407                                 $number = substr( $number, 1 );
3408                         }
3409                         $integerPart = [];
3410                         $decimalPart = [];
3411                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3412                         preg_match( "/\d+/", $number, $integerPart );
3413                         preg_match( "/\.\d*/", $number, $decimalPart );
3414                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3415                         if ( $groupedNumber === $number ) {
3416                                 // the string does not have any number part. Eg: .12345
3417                                 return $sign . $groupedNumber;
3418                         }
3419                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3420                         while ( $start > 0 ) {
3421                                 $match = $matches[0][$numMatches - 1];
3422                                 $matchLen = strlen( $match );
3423                                 $start = $end - $matchLen;
3424                                 if ( $start < 0 ) {
3425                                         $start = 0;
3426                                 }
3427                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3428                                 $end = $start;
3429                                 if ( $numMatches > 1 ) {
3430                                         // use the last pattern for the rest of the number
3431                                         $numMatches--;
3432                                 }
3433                                 if ( $start > 0 ) {
3434                                         $groupedNumber = "," . $groupedNumber;
3435                                 }
3436                         }
3437                         return $sign . $groupedNumber;
3438                 }
3439         }
3440
3441         /**
3442          * @return string
3443          */
3444         function digitGroupingPattern() {
3445                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3446         }
3447
3448         /**
3449          * @return array
3450          */
3451         function digitTransformTable() {
3452                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3453         }
3454
3455         /**
3456          * @return array
3457          */
3458         function separatorTransformTable() {
3459                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3460         }
3461
3462         /**
3463          * @return int|null
3464          */
3465         function minimumGroupingDigits() {
3466                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3467         }
3468
3469         /**
3470          * Take a list of strings and build a locale-friendly comma-separated
3471          * list, using the local comma-separator message.
3472          * The last two strings are chained with an "and".
3473          *
3474          * @param string[] $list
3475          * @return string
3476          */
3477         public function listToText( array $list ) {
3478                 $itemCount = count( $list );
3479                 if ( $itemCount < 1 ) {
3480                         return '';
3481                 }
3482                 $text = array_pop( $list );
3483                 if ( $itemCount > 1 ) {
3484                         $and = $this->msg( 'and' )->escaped();
3485                         $space = $this->msg( 'word-separator' )->escaped();
3486                         $comma = '';
3487                         if ( $itemCount > 2 ) {
3488                                 $comma = $this->msg( 'comma-separator' )->escaped();
3489                         }
3490                         $text = implode( $comma, $list ) . $and . $space . $text;
3491                 }
3492                 return $text;
3493         }
3494
3495         /**
3496          * Take a list of strings and build a locale-friendly comma-separated
3497          * list, using the local comma-separator message.
3498          * @param string[] $list Array of strings to put in a comma list
3499          * @return string
3500          */
3501         function commaList( array $list ) {
3502                 return implode(
3503                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3504                         $list
3505                 );
3506         }
3507
3508         /**
3509          * Take a list of strings and build a locale-friendly semicolon-separated
3510          * list, using the local semicolon-separator message.
3511          * @param string[] $list Array of strings to put in a semicolon list
3512          * @return string
3513          */
3514         function semicolonList( array $list ) {
3515                 return implode(
3516                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3517                         $list
3518                 );
3519         }
3520
3521         /**
3522          * Same as commaList, but separate it with the pipe instead.
3523          * @param string[] $list Array of strings to put in a pipe list
3524          * @return string
3525          */
3526         function pipeList( array $list ) {
3527                 return implode(
3528                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3529                         $list
3530                 );
3531         }
3532
3533         /**
3534          * This method is deprecated since 1.31 and kept as alias for truncateForDatabase, which
3535          * has replaced it. This method provides truncation suitable for DB.
3536          *
3537          * The database offers limited byte lengths for some columns in the database;
3538          * multi-byte character sets mean we need to ensure that only whole characters
3539          * are included, otherwise broken characters can be passed to the user.
3540          *
3541          * @deprecated since 1.31, use truncateForDatabase or truncateForVisual as appropriate.
3542          *
3543          * @param string $string String to truncate
3544          * @param int $length Maximum length (including ellipsis)
3545          * @param string $ellipsis String to append to the truncated text
3546          * @param bool $adjustLength Subtract length of ellipsis from $length.
3547          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3548          * @return string
3549          */
3550         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3551                 wfDeprecated( __METHOD__, '1.31' );
3552                 return $this->truncateForDatabase( $string, $length, $ellipsis, $adjustLength );
3553         }
3554
3555         /**
3556          * Truncate a string to a specified length in bytes, appending an optional
3557          * string (e.g. for ellipsis)
3558          *
3559          * If $length is negative, the string will be truncated from the beginning
3560          *
3561          * @since 1.31
3562          *
3563          * @param string $string String to truncate
3564          * @param int $length Maximum length in bytes
3565          * @param string $ellipsis String to append to the end of truncated text
3566          * @param bool $adjustLength Subtract length of ellipsis from $length
3567          *
3568          * @return string
3569          */
3570         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3571                 return $this->truncateInternal(
3572                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3573                 );
3574         }
3575
3576         /**
3577          * Truncate a string to a specified number of characters, appending an optional
3578          * string (e.g. for ellipsis).
3579          *
3580          * This provides multibyte version of truncate() method of this class, suitable for truncation
3581          * based on number of characters, instead of number of bytes.
3582          *
3583          * If $length is negative, the string will be truncated from the beginning.
3584          *
3585          * @since 1.31
3586          *
3587          * @param string $string String to truncate
3588          * @param int $length Maximum number of characters
3589          * @param string $ellipsis String to append to the end of truncated text
3590          * @param bool $adjustLength Subtract length of ellipsis from $length
3591          *
3592          * @return string
3593          */
3594         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3595                 // Passing encoding to mb_strlen and mb_substr is optional.
3596                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3597                 // explicit specification of encoding is skipped.
3598                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3599                 return $this->truncateInternal(
3600                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3601                 );
3602         }
3603
3604         /**
3605          * Internal method used for truncation. This method abstracts text truncation into
3606          * one common method, allowing users to provide length measurement function and
3607          * function for finding substring.
3608          *
3609          * For usages, see truncateForDatabase and truncateForVisual.
3610          *
3611          * @param string $string String to truncate
3612          * @param int $length Maximum length of final text
3613          * @param string $ellipsis String to append to the end of truncated text
3614          * @param bool $adjustLength Subtract length of ellipsis from $length
3615          * @param callable $measureLength Callable function used for determining the length of text
3616          * @param callable $getSubstring Callable function used for getting the substrings
3617          *
3618          * @return string
3619          */
3620         private function truncateInternal(
3621                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3622         ) {
3623                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3624                         throw new InvalidArgumentException( 'Invalid callback provided' );
3625                 }
3626
3627                 # Check if there is no need to truncate
3628                 if ( $measureLength( $string ) <= abs( $length ) ) {
3629                         return $string; // no need to truncate
3630                 }
3631
3632                 # Use the localized ellipsis character
3633                 if ( $ellipsis == '...' ) {
3634                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3635                 }
3636                 if ( $length == 0 ) {
3637                         return $ellipsis; // convention
3638                 }
3639
3640                 $stringOriginal = $string;
3641                 # If ellipsis length is >= $length then we can't apply $adjustLength
3642                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3643                         $string = $ellipsis; // this can be slightly unexpected
3644                 # Otherwise, truncate and add ellipsis...
3645                 } else {
3646                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3647                         if ( $length > 0 ) {
3648                                 $length -= $ellipsisLength;
3649                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3650                                 $string = $this->removeBadCharLast( $string );
3651                                 $string = rtrim( $string );
3652                                 $string = $string . $ellipsis;
3653                         } else {
3654                                 $length += $ellipsisLength;
3655                                 $string = $getSubstring( $string, $length ); // ...xyz
3656                                 $string = $this->removeBadCharFirst( $string );
3657                                 $string = ltrim( $string );
3658                                 $string = $ellipsis . $string;
3659                         }
3660                 }
3661
3662                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3663                 # This check is *not* redundant if $adjustLength, due to the single case where
3664                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3665                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3666                         return $string;
3667                 } else {
3668                         return $stringOriginal;
3669                 }
3670         }
3671
3672         /**
3673          * Remove bytes that represent an incomplete Unicode character
3674          * at the end of string (e.g. bytes of the char are missing)
3675          *
3676          * @param string $string
3677          * @return string
3678          */
3679         protected function removeBadCharLast( $string ) {
3680                 if ( $string != '' ) {
3681                         $char = ord( $string[strlen( $string ) - 1] );
3682                         $m = [];
3683                         if ( $char >= 0xc0 ) {
3684                                 # We got the first byte only of a multibyte char; remove it.
3685                                 $string = substr( $string, 0, -1 );
3686                         } elseif ( $char >= 0x80 &&
3687                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3688                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3689                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3690                         ) {
3691                                 # We chopped in the middle of a character; remove it
3692                                 $string = $m[1];
3693                         }
3694                 }
3695                 return $string;
3696         }
3697
3698         /**
3699          * Remove bytes that represent an incomplete Unicode character
3700          * at the start of string (e.g. bytes of the char are missing)
3701          *
3702          * @param string $string
3703          * @return string
3704          */
3705         protected function removeBadCharFirst( $string ) {
3706                 if ( $string != '' ) {
3707                         $char = ord( $string[0] );
3708                         if ( $char >= 0x80 && $char < 0xc0 ) {
3709                                 # We chopped in the middle of a character; remove the whole thing
3710                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3711                         }
3712                 }
3713                 return $string;
3714         }
3715
3716         /**
3717          * Truncate a string of valid HTML to a specified length in bytes,
3718          * appending an optional string (e.g. for ellipses), and return valid HTML
3719          *
3720          * This is only intended for styled/linked text, such as HTML with
3721          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3722          * Also, this will not detect things like "display:none" CSS.
3723          *
3724          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3725          *
3726          * @param string $text HTML string to truncate
3727          * @param int $length (zero/positive) Maximum length (including ellipses)
3728          * @param string $ellipsis String to append to the truncated text
3729          * @return string
3730          */
3731         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3732                 # Use the localized ellipsis character
3733                 if ( $ellipsis == '...' ) {
3734                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3735                 }
3736                 # Check if there is clearly no need to truncate
3737                 if ( $length <= 0 ) {
3738                         return $ellipsis; // no text shown, nothing to format (convention)
3739                 } elseif ( strlen( $text ) <= $length ) {
3740                         return $text; // string short enough even *with* HTML (short-circuit)
3741                 }
3742
3743                 $dispLen = 0; // innerHTML legth so far
3744                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3745                 $tagType = 0; // 0-open, 1-close
3746                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3747                 $entityState = 0; // 0-not entity, 1-entity
3748                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3749                 $openTags = []; // open tag stack
3750                 $maybeState = null; // possible truncation state
3751
3752                 $textLen = strlen( $text );
3753                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3754                 for ( $pos = 0; true; ++$pos ) {
3755                         # Consider truncation once the display length has reached the maximim.
3756                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3757                         # Check that we're not in the middle of a bracket/entity...
3758                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3759                                 if ( !$testingEllipsis ) {
3760                                         $testingEllipsis = true;
3761                                         # Save where we are; we will truncate here unless there turn out to
3762                                         # be so few remaining characters that truncation is not necessary.
3763                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3764                                                 $maybeState = [ $ret, $openTags ]; // save state
3765                                         }
3766                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3767                                         # String in fact does need truncation, the truncation point was OK.
3768                                         list( $ret, $openTags ) = $maybeState; // reload state
3769                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3770                                         $ret .= $ellipsis; // add ellipsis
3771                                         break;
3772                                 }
3773                         }
3774                         if ( $pos >= $textLen ) {
3775                                 break; // extra iteration just for above checks
3776                         }
3777
3778                         # Read the next char...
3779                         $ch = $text[$pos];
3780                         $lastCh = $pos ? $text[$pos - 1] : '';
3781                         $ret .= $ch; // add to result string
3782                         if ( $ch == '<' ) {
3783                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3784                                 $entityState = 0; // for bad HTML
3785                                 $bracketState = 1; // tag started (checking for backslash)
3786                         } elseif ( $ch == '>' ) {
3787                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3788                                 $entityState = 0; // for bad HTML
3789                                 $bracketState = 0; // out of brackets
3790                         } elseif ( $bracketState == 1 ) {
3791                                 if ( $ch == '/' ) {
3792                                         $tagType = 1; // close tag (e.g. "</span>")
3793                                 } else {
3794                                         $tagType = 0; // open tag (e.g. "<span>")
3795                                         $tag .= $ch;
3796                                 }
3797                                 $bracketState = 2; // building tag name
3798                         } elseif ( $bracketState == 2 ) {
3799                                 if ( $ch != ' ' ) {
3800                                         $tag .= $ch;
3801                                 } else {
3802                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3803                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3804                                 }
3805                         } elseif ( $bracketState == 0 ) {
3806                                 if ( $entityState ) {
3807                                         if ( $ch == ';' ) {
3808                                                 $entityState = 0;
3809                                                 $dispLen++; // entity is one displayed char
3810                                         }
3811                                 } else {
3812                                         if ( $neLength == 0 && !$maybeState ) {
3813                                                 // Save state without $ch. We want to *hit* the first
3814                                                 // display char (to get tags) but not *use* it if truncating.
3815                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3816                                         }
3817                                         if ( $ch == '&' ) {
3818                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3819                                         } else {
3820                                                 $dispLen++; // this char is displayed
3821                                                 // Add the next $max display text chars after this in one swoop...
3822                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3823                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3824                                                 $dispLen += $skipped;
3825                                                 $pos += $skipped;
3826                                         }
3827                                 }
3828                         }
3829                 }
3830                 // Close the last tag if left unclosed by bad HTML
3831                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3832                 while ( count( $openTags ) > 0 ) {
3833                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3834                 }
3835                 return $ret;
3836         }
3837
3838         /**
3839          * truncateHtml() helper function
3840          * like strcspn() but adds the skipped chars to $ret
3841          *
3842          * @param string $ret
3843          * @param string $text
3844          * @param string $search
3845          * @param int $start
3846          * @param null|int $len
3847          * @return int
3848          */
3849         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3850                 if ( $len === null ) {
3851                         $len = -1; // -1 means "no limit" for strcspn
3852                 } elseif ( $len < 0 ) {
3853                         $len = 0; // sanity
3854                 }
3855                 $skipCount = 0;
3856                 if ( $start < strlen( $text ) ) {
3857                         $skipCount = strcspn( $text, $search, $start, $len );
3858                         $ret .= substr( $text, $start, $skipCount );
3859                 }
3860                 return $skipCount;
3861         }
3862
3863         /**
3864          * truncateHtml() helper function
3865          * (a) push or pop $tag from $openTags as needed
3866          * (b) clear $tag value
3867          * @param string &$tag Current HTML tag name we are looking at
3868          * @param int $tagType (0-open tag, 1-close tag)
3869          * @param string $lastCh Character before the '>' that ended this tag
3870          * @param array &$openTags Open tag stack (not accounting for $tag)
3871          */
3872         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3873                 $tag = ltrim( $tag );
3874                 if ( $tag != '' ) {
3875                         if ( $tagType == 0 && $lastCh != '/' ) {
3876                                 $openTags[] = $tag; // tag opened (didn't close itself)
3877                         } elseif ( $tagType == 1 ) {
3878                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3879                                         array_pop( $openTags ); // tag closed
3880                                 }
3881                         }
3882                         $tag = '';
3883                 }
3884         }
3885
3886         /**
3887          * Grammatical transformations, needed for inflected languages
3888          * Invoked by putting {{grammar:case|word}} in a message
3889          *
3890          * @param string $word
3891          * @param string $case
3892          * @return string
3893          */
3894         function convertGrammar( $word, $case ) {
3895                 global $wgGrammarForms;
3896                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3897                         return $wgGrammarForms[$this->getCode()][$case][$word];
3898                 }
3899
3900                 $grammarTransformations = $this->getGrammarTransformations();
3901
3902                 if ( isset( $grammarTransformations[$case] ) ) {
3903                         $forms = $grammarTransformations[$case];
3904
3905                         // Some names of grammar rules are aliases for other rules.
3906                         // In such cases the value is a string rather than object,
3907                         // so load the actual rules.
3908                         if ( is_string( $forms ) ) {
3909                                 $forms = $grammarTransformations[$forms];
3910                         }
3911
3912                         foreach ( array_values( $forms ) as $rule ) {
3913                                 $form = $rule[0];
3914
3915                                 if ( $form === '@metadata' ) {
3916                                         continue;
3917                                 }
3918
3919                                 $replacement = $rule[1];
3920
3921                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3922                                 $patternMatches = preg_match( $regex, $word );
3923
3924                                 if ( $patternMatches === false ) {
3925                                         wfLogWarning(
3926                                                 'An error occurred while processing grammar. ' .
3927                                                 "Word: '$word'. Regex: /$form/."
3928                                         );
3929                                 } elseif ( $patternMatches === 1 ) {
3930                                         $word = preg_replace( $regex, $replacement, $word );
3931
3932                                         break;
3933                                 }
3934                         }
3935                 }
3936
3937                 return $word;
3938         }
3939
3940         /**
3941          * Get the grammar forms for the content language
3942          * @return array Array of grammar forms
3943          * @since 1.20
3944          */
3945         function getGrammarForms() {
3946                 global $wgGrammarForms;
3947                 if ( isset( $wgGrammarForms[$this->getCode()] )
3948                         && is_array( $wgGrammarForms[$this->getCode()] )
3949                 ) {
3950                         return $wgGrammarForms[$this->getCode()];
3951                 }
3952
3953                 return [];
3954         }
3955
3956         /**
3957          * Get the grammar transformations data for the language.
3958          * Used like grammar forms, with {{GRAMMAR}} and cases,
3959          * but uses pairs of regexes and replacements instead of code.
3960          *
3961          * @return array[] Array of grammar transformations.
3962          * @throws MWException
3963          * @since 1.28
3964          */
3965         public function getGrammarTransformations() {
3966                 $languageCode = $this->getCode();
3967
3968                 if ( self::$grammarTransformations === null ) {
3969                         self::$grammarTransformations = new MapCacheLRU( 10 );
3970                 }
3971
3972                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3973                         return self::$grammarTransformations->get( $languageCode );
3974                 }
3975
3976                 $data = [];
3977
3978                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3979                 if ( is_readable( $grammarDataFile ) ) {
3980                         $data = FormatJson::decode(
3981                                 file_get_contents( $grammarDataFile ),
3982                                 true
3983                         );
3984
3985                         if ( $data === null ) {
3986                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3987                         }
3988
3989                         self::$grammarTransformations->set( $languageCode, $data );
3990                 }
3991
3992                 return $data;
3993         }
3994
3995         /**
3996          * Provides an alternative text depending on specified gender.
3997          * Usage {{gender:username|masculine|feminine|unknown}}.
3998          * username is optional, in which case the gender of current user is used,
3999          * but only in (some) interface messages; otherwise default gender is used.
4000          *
4001          * If no forms are given, an empty string is returned. If only one form is
4002          * given, it will be returned unconditionally. These details are implied by
4003          * the caller and cannot be overridden in subclasses.
4004          *
4005          * If three forms are given, the default is to use the third (unknown) form.
4006          * If fewer than three forms are given, the default is to use the first (masculine) form.
4007          * These details can be overridden in subclasses.
4008          *
4009          * @param string $gender
4010          * @param array $forms
4011          *
4012          * @return string
4013          */
4014         function gender( $gender, $forms ) {
4015                 if ( !count( $forms ) ) {
4016                         return '';
4017                 }
4018                 $forms = $this->preConvertPlural( $forms, 2 );
4019                 if ( $gender === 'male' ) {
4020                         return $forms[0];
4021                 }
4022                 if ( $gender === 'female' ) {
4023                         return $forms[1];
4024                 }
4025                 return $forms[2] ?? $forms[0];
4026         }
4027
4028         /**
4029          * Plural form transformations, needed for some languages.
4030          * For example, there are 3 form of plural in Russian and Polish,
4031          * depending on "count mod 10". See [[w:Plural]]
4032          * For English it is pretty simple.
4033          *
4034          * Invoked by putting {{plural:count|wordform1|wordform2}}
4035          * or {{plural:count|wordform1|wordform2|wordform3}}
4036          *
4037          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
4038          *
4039          * @param int $count Non-localized number
4040          * @param array $forms Different plural forms
4041          * @return string Correct form of plural for $count in this language
4042          */
4043         function convertPlural( $count, $forms ) {
4044                 // Handle explicit n=pluralform cases
4045                 $forms = $this->handleExplicitPluralForms( $count, $forms );
4046                 if ( is_string( $forms ) ) {
4047                         return $forms;
4048                 }
4049                 if ( !count( $forms ) ) {
4050                         return '';
4051                 }
4052
4053                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
4054                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4055                 return $forms[$pluralForm];
4056         }
4057
4058         /**
4059          * Handles explicit plural forms for Language::convertPlural()
4060          *
4061          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4062          * If an explicitly defined plural form matches the $count, then
4063          * string value returned, otherwise array returned for further consideration
4064          * by CLDR rules or overridden convertPlural().
4065          *
4066          * @since 1.23
4067          *
4068          * @param int $count Non-localized number
4069          * @param array $forms Different plural forms
4070          *
4071          * @return array|string
4072          */
4073         protected function handleExplicitPluralForms( $count, array $forms ) {
4074                 foreach ( $forms as $index => $form ) {
4075                         if ( preg_match( '/\d+=/i', $form ) ) {
4076                                 $pos = strpos( $form, '=' );
4077                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4078                                         return substr( $form, $pos + 1 );
4079                                 }
4080                                 unset( $forms[$index] );
4081                         }
4082                 }
4083                 return array_values( $forms );
4084         }
4085
4086         /**
4087          * Checks that convertPlural was given an array and pads it to requested
4088          * amount of forms by copying the last one.
4089          *
4090          * @param array $forms Array of forms given to convertPlural
4091          * @param int $count How many forms should there be at least
4092          * @return array Padded array of forms or an exception if not an array
4093          */
4094         protected function preConvertPlural( /* Array */ $forms, $count ) {
4095                 while ( count( $forms ) < $count ) {
4096                         $forms[] = $forms[count( $forms ) - 1];
4097                 }
4098                 return $forms;
4099         }
4100
4101         /**
4102          * Wraps argument with unicode control characters for directionality safety
4103          *
4104          * This solves the problem where directionality-neutral characters at the edge of
4105          * the argument string get interpreted with the wrong directionality from the
4106          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4107          *
4108          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4109          * directionality of the argument string, using the BIDI algorithm's own "First
4110          * strong directional codepoint" rule. Essentially, this works round the fact that
4111          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4112          * direction inference). The latter is cleaner but still not widely supported.
4113          *
4114          * @param string $text Text to wrap
4115          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4116          */
4117         public function embedBidi( $text = '' ) {
4118                 $dir = self::strongDirFromContent( $text );
4119                 if ( $dir === 'ltr' ) {
4120                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4121                         return self::$lre . $text . self::$pdf;
4122                 }
4123                 if ( $dir === 'rtl' ) {
4124                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4125                         return self::$rle . $text . self::$pdf;
4126                 }
4127                 // No strong directionality: do not wrap
4128                 return $text;
4129         }
4130
4131         /**
4132          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4133          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4134          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4135          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4136          * match up with it.
4137          *
4138          * @param string $str The validated block duration in English
4139          * @param User|null $user User object to use timezone from or null for $wgUser
4140          * @param int $now Current timestamp, for formatting relative block durations
4141          * @return string Somehow translated block duration
4142          * @see LanguageFi.php for example implementation
4143          */
4144         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4145                 $duration = SpecialBlock::getSuggestedDurations( $this );
4146                 foreach ( $duration as $show => $value ) {
4147                         if ( strcmp( $str, $value ) == 0 ) {
4148                                 return htmlspecialchars( trim( $show ) );
4149                         }
4150                 }
4151
4152                 if ( wfIsInfinity( $str ) ) {
4153                         foreach ( $duration as $show => $value ) {
4154                                 if ( wfIsInfinity( $value ) ) {
4155                                         return htmlspecialchars( trim( $show ) );
4156                                 }
4157                         }
4158                 }
4159
4160                 // If all else fails, return a standard duration or timestamp description.
4161                 $time = strtotime( $str, $now );
4162                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4163                         return $str;
4164                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4165                         // The result differs based on current time, so the difference
4166                         // is a fixed duration length.
4167                         return $this->formatDuration( $time - $now );
4168                 } else { // It's an absolute timestamp.
4169                         if ( $time === 0 ) {
4170                                 // wfTimestamp() handles 0 as current time instead of epoch.
4171                                 $time = '19700101000000';
4172                         }
4173                         if ( $user ) {
4174                                 return $this->userTimeAndDate( $time, $user );
4175                         }
4176                         return $this->timeanddate( $time );
4177                 }
4178         }
4179
4180         /**
4181          * languages like Chinese need to be segmented in order for the diff
4182          * to be of any use
4183          *
4184          * @param string $text
4185          * @return string
4186          */
4187         public function segmentForDiff( $text ) {
4188                 return $text;
4189         }
4190
4191         /**
4192          * and unsegment to show the result
4193          *
4194          * @param string $text
4195          * @return string
4196          */
4197         public function unsegmentForDiff( $text ) {
4198                 return $text;
4199         }
4200
4201         /**
4202          * Return the LanguageConverter used in the Language
4203          *
4204          * @since 1.19
4205          * @return LanguageConverter
4206          */
4207         public function getConverter() {
4208                 return $this->mConverter;
4209         }
4210
4211         /**
4212          * convert text to a variant
4213          *
4214          * @param string $text text to convert
4215          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4216          *      variant (if logged in), or the project default variant
4217          * @return string the converted string
4218          */
4219         public function autoConvert( $text, $variant = false ) {
4220                 return $this->mConverter->autoConvert( $text, $variant );
4221         }
4222
4223         /**
4224          * convert text to all supported variants
4225          *
4226          * @param string $text
4227          * @return array
4228          */
4229         public function autoConvertToAllVariants( $text ) {
4230                 return $this->mConverter->autoConvertToAllVariants( $text );
4231         }
4232
4233         /**
4234          * convert text to different variants of a language.
4235          *
4236          * @warning Glossary state is maintained between calls. This means
4237          *  if you pass unescaped text to this method it can cause an XSS
4238          *  in later calls to this method, even if the later calls have properly
4239          *  escaped the input. Never feed this method user controlled text that
4240          *  is not properly escaped!
4241          * @param string $text Content that has been already escaped for use in HTML
4242          * @return string HTML
4243          */
4244         public function convert( $text ) {
4245                 return $this->mConverter->convert( $text );
4246         }
4247
4248         /**
4249          * Convert a Title object to a string in the preferred variant
4250          *
4251          * @param Title $title
4252          * @return string
4253          */
4254         public function convertTitle( $title ) {
4255                 return $this->mConverter->convertTitle( $title );
4256         }
4257
4258         /**
4259          * Convert a namespace index to a string in the preferred variant
4260          *
4261          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4262          * @param string|null $variant variant to convert to, or null to use the user's preferred
4263          *      variant (if logged in), or the project default variant
4264          * @return string a string representation of the namespace
4265          */
4266         public function convertNamespace( $ns, $variant = null ) {
4267                 return $this->mConverter->convertNamespace( $ns, $variant );
4268         }
4269
4270         /**
4271          * Check if this is a language with variants
4272          *
4273          * @return bool
4274          */
4275         public function hasVariants() {
4276                 return count( $this->getVariants() ) > 1;
4277         }
4278
4279         /**
4280          * Check if the language has the specific variant
4281          *
4282          * @since 1.19
4283          * @param string $variant
4284          * @return bool
4285          */
4286         public function hasVariant( $variant ) {
4287                 return (bool)$this->mConverter->validateVariant( $variant );
4288         }
4289
4290         /**
4291          * Perform output conversion on a string, and encode for safe HTML output.
4292          * @param string $text Text to be converted
4293          * @param bool $isTitle Whether this conversion is for the article title
4294          * @return string
4295          * @todo this should get integrated somewhere sane
4296          */
4297         public function convertHtml( $text, $isTitle = false ) {
4298                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4299         }
4300
4301         /**
4302          * @param string $key
4303          * @return string
4304          */
4305         public function convertCategoryKey( $key ) {
4306                 return $this->mConverter->convertCategoryKey( $key );
4307         }
4308
4309         /**
4310          * Get the list of variants supported by this language
4311          * see sample implementation in LanguageZh.php
4312          *
4313          * @return string[] An array of language codes
4314          */
4315         public function getVariants() {
4316                 return $this->mConverter->getVariants();
4317         }
4318
4319         /**
4320          * @return string
4321          */
4322         public function getPreferredVariant() {
4323                 return $this->mConverter->getPreferredVariant();
4324         }
4325
4326         /**
4327          * @return string
4328          */
4329         public function getDefaultVariant() {
4330                 return $this->mConverter->getDefaultVariant();
4331         }
4332
4333         /**
4334          * @return string
4335          */
4336         public function getURLVariant() {
4337                 return $this->mConverter->getURLVariant();
4338         }
4339
4340         /**
4341          * If a language supports multiple variants, it is
4342          * possible that non-existing link in one variant
4343          * actually exists in another variant. this function
4344          * tries to find it. See e.g. LanguageZh.php
4345          * The input parameters may be modified upon return
4346          *
4347          * @param string &$link The name of the link
4348          * @param Title &$nt The title object of the link
4349          * @param bool $ignoreOtherCond To disable other conditions when
4350          *   we need to transclude a template or update a category's link
4351          */
4352         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4353                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4354         }
4355
4356         /**
4357          * returns language specific options used by User::getPageRenderHash()
4358          * for example, the preferred language variant
4359          *
4360          * @return string
4361          */
4362         function getExtraHashOptions() {
4363                 return $this->mConverter->getExtraHashOptions();
4364         }
4365
4366         /**
4367          * For languages that support multiple variants, the title of an
4368          * article may be displayed differently in different variants. this
4369          * function returns the apporiate title defined in the body of the article.
4370          *
4371          * @return string
4372          */
4373         public function getParsedTitle() {
4374                 return $this->mConverter->getParsedTitle();
4375         }
4376
4377         /**
4378          * Refresh the cache of conversion tables when
4379          * MediaWiki:Conversiontable* is updated.
4380          *
4381          * @param Title $title The Title of the page being updated
4382          */
4383         public function updateConversionTable( Title $title ) {
4384                 $this->mConverter->updateConversionTable( $title );
4385         }
4386
4387         /**
4388          * Prepare external link text for conversion. When the text is
4389          * a URL, it shouldn't be converted, and it'll be wrapped in
4390          * the "raw" tag (-{R| }-) to prevent conversion.
4391          *
4392          * This function is called "markNoConversion" for historical
4393          * reasons *BUT DIFFERS SIGNIFICANTLY* from
4394          * LanguageConverter::markNoConversion(), with which it is easily
4395          * confused.
4396          *
4397          * @param string $text Text to be used for external link
4398          * @param bool $noParse Wrap it without confirming it's a real URL first
4399          * @return string The tagged text
4400          * @deprecated since 1.32, use LanguageConverter::markNoConversion()
4401          *  instead.
4402          */
4403         public function markNoConversion( $text, $noParse = false ) {
4404                 wfDeprecated( __METHOD__, '1.32' );
4405                 // Excluding protocal-relative URLs may avoid many false positives.
4406                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4407                         return $this->mConverter->markNoConversion( $text );
4408                 } else {
4409                         return $text;
4410                 }
4411         }
4412
4413         /**
4414          * A regular expression to match legal word-trailing characters
4415          * which should be merged onto a link of the form [[foo]]bar.
4416          *
4417          * @return string
4418          */
4419         public function linkTrail() {
4420                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4421         }
4422
4423         /**
4424          * A regular expression character set to match legal word-prefixing
4425          * characters which should be merged onto a link of the form foo[[bar]].
4426          *
4427          * @return string
4428          */
4429         public function linkPrefixCharset() {
4430                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4431         }
4432
4433         /**
4434          * Get the "parent" language which has a converter to convert a "compatible" language
4435          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4436          *
4437          * @return Language|null
4438          * @since 1.22
4439          */
4440         public function getParentLanguage() {
4441                 if ( $this->mParentLanguage !== false ) {
4442                         return $this->mParentLanguage;
4443                 }
4444
4445                 $code = explode( '-', $this->getCode() )[0];
4446                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4447                         $this->mParentLanguage = null;
4448                         return null;
4449                 }
4450                 $lang = self::factory( $code );
4451                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4452                         $this->mParentLanguage = null;
4453                         return null;
4454                 }
4455
4456                 $this->mParentLanguage = $lang;
4457                 return $lang;
4458         }
4459
4460         /**
4461          * Compare with an other language object
4462          *
4463          * @since 1.28
4464          * @param Language $lang
4465          * @return bool
4466          */
4467         public function equals( Language $lang ) {
4468                 return $lang === $this || $lang->getCode() === $this->mCode;
4469         }
4470
4471         /**
4472          * Get the internal language code for this language object
4473          *
4474          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4475          * htmlspecialchars() or similar
4476          *
4477          * @return string
4478          */
4479         public function getCode() {
4480                 return $this->mCode;
4481         }
4482
4483         /**
4484          * Get the code in BCP 47 format which we can use
4485          * inside of html lang="" tags.
4486          *
4487          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4488          * htmlspecialchars() or similar.
4489          *
4490          * @since 1.19
4491          * @return string
4492          */
4493         public function getHtmlCode() {
4494                 if ( is_null( $this->mHtmlCode ) ) {
4495                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4496                 }
4497                 return $this->mHtmlCode;
4498         }
4499
4500         /**
4501          * @param string $code
4502          */
4503         public function setCode( $code ) {
4504                 $this->mCode = $code;
4505                 // Ensure we don't leave incorrect cached data lying around
4506                 $this->mHtmlCode = null;
4507                 $this->mParentLanguage = false;
4508         }
4509
4510         /**
4511          * Get the language code from a file name. Inverse of getFileName()
4512          * @param string $filename $prefix . $languageCode . $suffix
4513          * @param string $prefix Prefix before the language code
4514          * @param string $suffix Suffix after the language code
4515          * @return string Language code, or false if $prefix or $suffix isn't found
4516          */
4517         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4518                 $m = null;
4519                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4520                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4521                 if ( !count( $m ) ) {
4522                         return false;
4523                 }
4524                 return str_replace( '_', '-', strtolower( $m[1] ) );
4525         }
4526
4527         /**
4528          * @param string $code
4529          * @param bool $fallback Whether we're going through language fallback chain
4530          * @return string Name of the language class
4531          */
4532         public static function classFromCode( $code, $fallback = true ) {
4533                 if ( $fallback && $code == 'en' ) {
4534                         return 'Language';
4535                 } else {
4536                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4537                 }
4538         }
4539
4540         /**
4541          * Get the name of a file for a certain language code
4542          * @param string $prefix Prepend this to the filename
4543          * @param string $code Language code
4544          * @param string $suffix Append this to the filename
4545          * @throws MWException
4546          * @return string $prefix . $mangledCode . $suffix
4547          */
4548         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4549                 if ( !self::isValidBuiltInCode( $code ) ) {
4550                         throw new MWException( "Invalid language code \"$code\"" );
4551                 }
4552
4553                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4554         }
4555
4556         /**
4557          * @param string $code
4558          * @return string
4559          */
4560         public static function getMessagesFileName( $code ) {
4561                 global $IP;
4562                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4563                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4564                 return $file;
4565         }
4566
4567         /**
4568          * @param string $code
4569          * @return string
4570          * @throws MWException
4571          * @since 1.23
4572          */
4573         public static function getJsonMessagesFileName( $code ) {
4574                 global $IP;
4575
4576                 if ( !self::isValidBuiltInCode( $code ) ) {
4577                         throw new MWException( "Invalid language code \"$code\"" );
4578                 }
4579
4580                 return "$IP/languages/i18n/$code.json";
4581         }
4582
4583         /**
4584          * Get the first fallback for a given language.
4585          *
4586          * @param string $code
4587          *
4588          * @return bool|string
4589          */
4590         public static function getFallbackFor( $code ) {
4591                 $fallbacks = self::getFallbacksFor( $code );
4592                 if ( $fallbacks ) {
4593                         return $fallbacks[0];
4594                 }
4595                 return false;
4596         }
4597
4598         /**
4599          * Get the ordered list of fallback languages.
4600          *
4601          * @since 1.19
4602          * @param string $code Language code
4603          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4604          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4605          * @throws MWException
4606          * @return array List of language codes
4607          */
4608         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4609                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4610                         return [];
4611                 }
4612                 switch ( $mode ) {
4613                         case self::MESSAGES_FALLBACKS:
4614                                 // For unknown languages, fallbackSequence returns an empty array,
4615                                 // hardcode fallback to 'en' in that case as English messages are
4616                                 // always defined.
4617                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4618                         case self::STRICT_FALLBACKS:
4619                                 // Use this mode when you don't want to fallback to English unless
4620                                 // explicitly defined, for example when you have language-variant icons
4621                                 // and an international language-independent fallback.
4622                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4623                         default:
4624                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4625                 }
4626         }
4627
4628         /**
4629          * Get the ordered list of fallback languages, ending with the fallback
4630          * language chain for the site language.
4631          *
4632          * @since 1.22
4633          * @param string $code Language code
4634          * @return array Array( fallbacks, site fallbacks )
4635          */
4636         public static function getFallbacksIncludingSiteLanguage( $code ) {
4637                 global $wgLanguageCode;
4638
4639                 // Usually, we will only store a tiny number of fallback chains, so we
4640                 // keep them in static memory.
4641                 $cacheKey = "{$code}-{$wgLanguageCode}";
4642
4643                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4644                         $fallbacks = self::getFallbacksFor( $code );
4645
4646                         // Append the site's fallback chain, including the site language itself
4647                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4648                         array_unshift( $siteFallbacks, $wgLanguageCode );
4649
4650                         // Eliminate any languages already included in the chain
4651                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4652
4653                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4654                 }
4655                 return self::$fallbackLanguageCache[$cacheKey];
4656         }
4657
4658         /**
4659          * Get all messages for a given language
4660          * WARNING: this may take a long time. If you just need all message *keys*
4661          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4662          *
4663          * @param string $code
4664          *
4665          * @return array
4666          */
4667         public static function getMessagesFor( $code ) {
4668                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4669         }
4670
4671         /**
4672          * Get a message for a given language
4673          *
4674          * @param string $key
4675          * @param string $code
4676          *
4677          * @return string
4678          */
4679         public static function getMessageFor( $key, $code ) {
4680                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4681         }
4682
4683         /**
4684          * Get all message keys for a given language. This is a faster alternative to
4685          * array_keys( Language::getMessagesFor( $code ) )
4686          *
4687          * @since 1.19
4688          * @param string $code Language code
4689          * @return array Array of message keys (strings)
4690          */
4691         public static function getMessageKeysFor( $code ) {
4692                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4693         }
4694
4695         /**
4696          * @param string $talk
4697          * @return mixed
4698          */
4699         function fixVariableInNamespace( $talk ) {
4700                 if ( strpos( $talk, '$1' ) === false ) {
4701                         return $talk;
4702                 }
4703
4704                 global $wgMetaNamespace;
4705                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4706
4707                 # Allow grammar transformations
4708                 # Allowing full message-style parsing would make simple requests
4709                 # such as action=raw much more expensive than they need to be.
4710                 # This will hopefully cover most cases.
4711                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4712                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4713                 return str_replace( ' ', '_', $talk );
4714         }
4715
4716         /**
4717          * @param string $m
4718          * @return string
4719          */
4720         function replaceGrammarInNamespace( $m ) {
4721                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4722         }
4723
4724         /**
4725          * Decode an expiry (block, protection, etc) which has come from the DB
4726          *
4727          * @param string $expiry Database expiry String
4728          * @param bool|int $format True to process using language functions, or TS_ constant
4729          *     to return the expiry in a given timestamp
4730          * @param string $infinity If $format is not true, use this string for infinite expiry
4731          * @return string
4732          * @since 1.18
4733          */
4734         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4735                 static $dbInfinity;
4736                 if ( $dbInfinity === null ) {
4737                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4738                 }
4739
4740                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4741                         return $format === true
4742                                 ? $this->getMessageFromDB( 'infiniteblock' )
4743                                 : $infinity;
4744                 } else {
4745                         return $format === true
4746                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4747                                 : wfTimestamp( $format, $expiry );
4748                 }
4749         }
4750
4751         /**
4752          * Formats a time given in seconds into a string representation of that time.
4753          *
4754          * @param int|float $seconds
4755          * @param array $format An optional argument that formats the returned string in different ways:
4756          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4757          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4758          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4759          *     and friends.
4760          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4761          *     or 'avoidminutes'.
4762          * @return string
4763          */
4764         function formatTimePeriod( $seconds, $format = [] ) {
4765                 if ( !is_array( $format ) ) {
4766                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4767                 }
4768                 if ( !isset( $format['avoid'] ) ) {
4769                         $format['avoid'] = false;
4770                 }
4771                 if ( !isset( $format['noabbrevs'] ) ) {
4772                         $format['noabbrevs'] = false;
4773                 }
4774                 $secondsMsg = wfMessage(
4775                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4776                 $minutesMsg = wfMessage(
4777                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4778                 $hoursMsg = wfMessage(
4779                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4780                 $daysMsg = wfMessage(
4781                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4782
4783                 if ( round( $seconds * 10 ) < 100 ) {
4784                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4785                         $s = $secondsMsg->params( $s )->text();
4786                 } elseif ( round( $seconds ) < 60 ) {
4787                         $s = $this->formatNum( round( $seconds ) );
4788                         $s = $secondsMsg->params( $s )->text();
4789                 } elseif ( round( $seconds ) < 3600 ) {
4790                         $minutes = floor( $seconds / 60 );
4791                         $secondsPart = round( fmod( $seconds, 60 ) );
4792                         if ( $secondsPart == 60 ) {
4793                                 $secondsPart = 0;
4794                                 $minutes++;
4795                         }
4796                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4797                         $s .= ' ';
4798                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4799                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4800                         $hours = floor( $seconds / 3600 );
4801                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4802                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4803                         if ( $secondsPart == 60 ) {
4804                                 $secondsPart = 0;
4805                                 $minutes++;
4806                         }
4807                         if ( $minutes == 60 ) {
4808                                 $minutes = 0;
4809                                 $hours++;
4810                         }
4811                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4812                         $s .= ' ';
4813                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4814                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4815                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4816                         }
4817                 } else {
4818                         $days = floor( $seconds / 86400 );
4819                         if ( $format['avoid'] === 'avoidminutes' ) {
4820                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4821                                 if ( $hours == 24 ) {
4822                                         $hours = 0;
4823                                         $days++;
4824                                 }
4825                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4826                                 $s .= ' ';
4827                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4828                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4829                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4830                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4831                                 if ( $minutes == 60 ) {
4832                                         $minutes = 0;
4833                                         $hours++;
4834                                 }
4835                                 if ( $hours == 24 ) {
4836                                         $hours = 0;
4837                                         $days++;
4838                                 }
4839                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4840                                 $s .= ' ';
4841                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4842                                 $s .= ' ';
4843                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4844                         } else {
4845                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4846                                 $s .= ' ';
4847                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4848                         }
4849                 }
4850                 return $s;
4851         }
4852
4853         /**
4854          * Format a bitrate for output, using an appropriate
4855          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4856          *   the magnitude in question.
4857          *
4858          * This use base 1000. For base 1024 use formatSize(), for another base
4859          * see formatComputingNumbers().
4860          *
4861          * @param int $bps
4862          * @return string
4863          */
4864         function formatBitrate( $bps ) {
4865                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4866         }
4867
4868         /**
4869          * @param int $size Size of the unit
4870          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4871          * @param string $messageKey Message key to be uesd
4872          * @return string
4873          */
4874         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4875                 if ( $size <= 0 ) {
4876                         return str_replace( '$1', $this->formatNum( $size ),
4877                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4878                         );
4879                 }
4880                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4881                 $index = 0;
4882
4883                 $maxIndex = count( $sizes ) - 1;
4884                 while ( $size >= $boundary && $index < $maxIndex ) {
4885                         $index++;
4886                         $size /= $boundary;
4887                 }
4888
4889                 // For small sizes no decimal places necessary
4890                 $round = 0;
4891                 if ( $index > 1 ) {
4892                         // For MB and bigger two decimal places are smarter
4893                         $round = 2;
4894                 }
4895                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4896
4897                 $size = round( $size, $round );
4898                 $text = $this->getMessageFromDB( $msg );
4899                 return str_replace( '$1', $this->formatNum( $size ), $text );
4900         }
4901
4902         /**
4903          * Format a size in bytes for output, using an appropriate
4904          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4905          *
4906          * This method use base 1024. For base 1000 use formatBitrate(), for
4907          * another base see formatComputingNumbers()
4908          *
4909          * @param int $size Size to format
4910          * @return string Plain text (not HTML)
4911          */
4912         function formatSize( $size ) {
4913                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4914         }
4915
4916         /**
4917          * Make a list item, used by various special pages
4918          *
4919          * @param string $page Page link
4920          * @param string $details HTML safe text between brackets
4921          * @param bool $oppositedm Add the direction mark opposite to your
4922          *   language, to display text properly
4923          * @return string HTML escaped
4924          */
4925         function specialList( $page, $details, $oppositedm = true ) {
4926                 if ( !$details ) {
4927                         return $page;
4928                 }
4929
4930                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4931                 return $page .
4932                         $dirmark .
4933                         $this->msg( 'word-separator' )->escaped() .
4934                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4935         }
4936
4937         /**
4938          * Generate (prev x| next x) (20|50|100...) type links for paging
4939          *
4940          * @param Title $title Title object to link
4941          * @param int $offset
4942          * @param int $limit
4943          * @param array $query Optional URL query parameter string
4944          * @param bool $atend Optional param for specified if this is the last page
4945          * @return string
4946          */
4947         public function viewPrevNext( Title $title, $offset, $limit,
4948                 array $query = [], $atend = false
4949         ) {
4950                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4951
4952                 # Make 'previous' link
4953                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4954                 if ( $offset > 0 ) {
4955                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4956                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4957                 } else {
4958                         $plink = htmlspecialchars( $prev );
4959                 }
4960
4961                 # Make 'next' link
4962                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4963                 if ( $atend ) {
4964                         $nlink = htmlspecialchars( $next );
4965                 } else {
4966                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4967                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4968                 }
4969
4970                 # Make links to set number of items per page
4971                 $numLinks = [];
4972                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4973                         $numLinks[] = $this->numLink( $title, $offset, $num,
4974                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4975                 }
4976
4977                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4978                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4979         }
4980
4981         /**
4982          * Helper function for viewPrevNext() that generates links
4983          *
4984          * @param Title $title Title object to link
4985          * @param int $offset
4986          * @param int $limit
4987          * @param array $query Extra query parameters
4988          * @param string $link Text to use for the link; will be escaped
4989          * @param string $tooltipMsg Name of the message to use as tooltip
4990          * @param string $class Value of the "class" attribute of the link
4991          * @return string HTML fragment
4992          */
4993         private function numLink( Title $title, $offset, $limit, array $query, $link,
4994                 $tooltipMsg, $class
4995         ) {
4996                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4997                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4998                         ->numParams( $limit )->text();
4999
5000                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
5001                         'title' => $tooltip, 'class' => $class ], $link );
5002         }
5003
5004         /**
5005          * Get the conversion rule title, if any.
5006          *
5007          * @return string
5008          */
5009         public function getConvRuleTitle() {
5010                 return $this->mConverter->getConvRuleTitle();
5011         }
5012
5013         /**
5014          * Get the compiled plural rules for the language
5015          * @since 1.20
5016          * @return array Associative array with plural form, and plural rule as key-value pairs
5017          */
5018         public function getCompiledPluralRules() {
5019                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
5020                 $fallbacks = self::getFallbacksFor( $this->mCode );
5021                 if ( !$pluralRules ) {
5022                         foreach ( $fallbacks as $fallbackCode ) {
5023                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
5024                                 if ( $pluralRules ) {
5025                                         break;
5026                                 }
5027                         }
5028                 }
5029                 return $pluralRules;
5030         }
5031
5032         /**
5033          * Get the plural rules for the language
5034          * @since 1.20
5035          * @return array Associative array with plural form number and plural rule as key-value pairs
5036          */
5037         public function getPluralRules() {
5038                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
5039                 $fallbacks = self::getFallbacksFor( $this->mCode );
5040                 if ( !$pluralRules ) {
5041                         foreach ( $fallbacks as $fallbackCode ) {
5042                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
5043                                 if ( $pluralRules ) {
5044                                         break;
5045                                 }
5046                         }
5047                 }
5048                 return $pluralRules;
5049         }
5050
5051         /**
5052          * Get the plural rule types for the language
5053          * @since 1.22
5054          * @return array Associative array with plural form number and plural rule type as key-value pairs
5055          */
5056         public function getPluralRuleTypes() {
5057                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
5058                 $fallbacks = self::getFallbacksFor( $this->mCode );
5059                 if ( !$pluralRuleTypes ) {
5060                         foreach ( $fallbacks as $fallbackCode ) {
5061                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
5062                                 if ( $pluralRuleTypes ) {
5063                                         break;
5064                                 }
5065                         }
5066                 }
5067                 return $pluralRuleTypes;
5068         }
5069
5070         /**
5071          * Find the index number of the plural rule appropriate for the given number
5072          * @param int $number
5073          * @return int The index number of the plural rule
5074          */
5075         public function getPluralRuleIndexNumber( $number ) {
5076                 $pluralRules = $this->getCompiledPluralRules();
5077                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5078                 return $form;
5079         }
5080
5081         /**
5082          * Find the plural rule type appropriate for the given number
5083          * For example, if the language is set to Arabic, getPluralType(5) should
5084          * return 'few'.
5085          * @since 1.22
5086          * @param int $number
5087          * @return string The name of the plural rule type, e.g. one, two, few, many
5088          */
5089         public function getPluralRuleType( $number ) {
5090                 $index = $this->getPluralRuleIndexNumber( $number );
5091                 $pluralRuleTypes = $this->getPluralRuleTypes();
5092                 if ( isset( $pluralRuleTypes[$index] ) ) {
5093                         return $pluralRuleTypes[$index];
5094                 } else {
5095                         return 'other';
5096                 }
5097         }
5098 }