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 = [];
  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 string[] List of localized namespace names, indexed by numeric namespace ID.
 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                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 708                         # that aren't loaded here but are included in the l10n cache.
 709                         # (array_intersect preserves keys from its first argument)
 710                         $this->namespaceAliases = array_intersect(
 711                                 $this->namespaceAliases,
 712                                 array_keys( $this->getNamespaces() )
 713                         );
 714                 }
 715
 716                 return $this->namespaceAliases;
 717         }
 718
 719         /**
 720          * @return array
 721          */
 722         public function getNamespaceIds() {
 723                 if ( is_null( $this->mNamespaceIds ) ) {
 724                         global $wgNamespaceAliases;
 725                         # Put namespace names and aliases into a hashtable.
 726                         # If this is too slow, then we should arrange it so that it is done
 727                         # before caching. The catch is that at pre-cache time, the above
 728                         # class-specific fixup hasn't been done.
 729                         $this->mNamespaceIds = [];
 730                         foreach ( $this->getNamespaces() as $index => $name ) {
 731                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 732                         }
 733                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 734                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 735                         }
 736                         if ( $wgNamespaceAliases ) {
 737                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 738                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 739                                 }
 740                         }
 741                 }
 742                 return $this->mNamespaceIds;
 743         }
 744
 745         /**
 746          * Get a namespace key by value, case insensitive.  Canonical namespace
 747          * names override custom ones defined for the current language.
 748          *
 749          * @param string $text
 750          * @return int|bool An integer if $text is a valid value otherwise false
 751          */
 752         public function getNsIndex( $text ) {
 753                 $lctext = $this->lc( $text );
 754                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 755                 if ( $ns !== null ) {
 756                         return $ns;
 757                 }
 758                 $ids = $this->getNamespaceIds();
 759                 return $ids[$lctext] ?? false;
 760         }
 761
 762         /**
 763          * short names for language variants used for language conversion links.
 764          *
 765          * @param string $code
 766          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 767          * @return string
 768          */
 769         public function getVariantname( $code, $usemsg = true ) {
 770                 $msg = "variantname-$code";
 771                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 772                         return $this->getMessageFromDB( $msg );
 773                 }
 774                 $name = self::fetchLanguageName( $code );
 775                 if ( $name ) {
 776                         return $name; # if it's defined as a language name, show that
 777                 } else {
 778                         # otherwise, output the language code
 779                         return $code;
 780                 }
 781         }
 782
 783         /**
 784          * @return string[]|bool List of date format preference keys, or false if disabled.
 785          */
 786         public function getDatePreferences() {
 787                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 788         }
 789
 790         /**
 791          * @return array
 792          */
 793         function getDateFormats() {
 794                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 795         }
 796
 797         /**
 798          * @return array|string
 799          */
 800         public function getDefaultDateFormat() {
 801                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 802                 if ( $df === 'dmy or mdy' ) {
 803                         global $wgAmericanDates;
 804                         return $wgAmericanDates ? 'mdy' : 'dmy';
 805                 } else {
 806                         return $df;
 807                 }
 808         }
 809
 810         /**
 811          * @return array
 812          */
 813         public function getDatePreferenceMigrationMap() {
 814                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 815         }
 816
 817         /**
 818          * @return array
 819          */
 820         public function getExtraUserToggles() {
 821                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 822         }
 823
 824         /**
 825          * @param string $tog
 826          * @return string
 827          */
 828         function getUserToggle( $tog ) {
 829                 return $this->getMessageFromDB( "tog-$tog" );
 830         }
 831
 832         /**
 833          * Get an array of language names, indexed by code.
 834          * @param null|string $inLanguage Code of language in which to return the names
 835          *              Use self::AS_AUTONYMS for autonyms (native names)
 836          * @param string $include One of:
 837          *              self::ALL all available languages
 838          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 839          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 840          * @return array Language code => language name (sorted by key)
 841          * @since 1.20
 842          */
 843         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 844                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 845                 $cacheKey .= ":$include";
 846                 if ( self::$languageNameCache === null ) {
 847                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 848                 }
 849
 850                 $ret = self::$languageNameCache->get( $cacheKey );
 851                 if ( !$ret ) {
 852                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 853                         self::$languageNameCache->set( $cacheKey, $ret );
 854                 }
 855                 return $ret;
 856         }
 857
 858         /**
 859          * Uncached helper for fetchLanguageNames
 860          * @param null|string $inLanguage Code of language in which to return the names
 861          *              Use self::AS_AUTONYMS for autonyms (native names)
 862          * @param string $include One of:
 863          *              self::ALL all available languages
 864          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 865          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 866          * @return array Language code => language name (sorted by key)
 867          */
 868         private static function fetchLanguageNamesUncached(
 869                 $inLanguage = self::AS_AUTONYMS,
 870                 $include = 'mw'
 871         ) {
 872                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 873
 874                 // If passed an invalid language code to use, fallback to en
 875                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 876                         $inLanguage = 'en';
 877                 }
 878
 879                 $names = [];
 880
 881                 if ( $inLanguage ) {
 882                         # TODO: also include when $inLanguage is null, when this code is more efficient
 883                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 884                 }
 885
 886                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 887                 if ( $wgUsePigLatinVariant ) {
 888                         // Pig Latin (for variant development)
 889                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 890                 }
 891
 892                 foreach ( $mwNames as $mwCode => $mwName ) {
 893                         # - Prefer own MediaWiki native name when not using the hook
 894                         # - For other names just add if not added through the hook
 895                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 896                                 $names[$mwCode] = $mwName;
 897                         }
 898                 }
 899
 900                 if ( $include === self::ALL ) {
 901                         ksort( $names );
 902                         return $names;
 903                 }
 904
 905                 $returnMw = [];
 906                 $coreCodes = array_keys( $mwNames );
 907                 foreach ( $coreCodes as $coreCode ) {
 908                         $returnMw[$coreCode] = $names[$coreCode];
 909                 }
 910
 911                 if ( $include === self::SUPPORTED ) {
 912                         $namesMwFile = [];
 913                         # We do this using a foreach over the codes instead of a directory
 914                         # loop so that messages files in extensions will work correctly.
 915                         foreach ( $returnMw as $code => $value ) {
 916                                 if ( is_readable( self::getMessagesFileName( $code ) )
 917                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 918                                 ) {
 919                                         $namesMwFile[$code] = $names[$code];
 920                                 }
 921                         }
 922
 923                         ksort( $namesMwFile );
 924                         return $namesMwFile;
 925                 }
 926
 927                 ksort( $returnMw );
 928                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 929                 return $returnMw;
 930         }
 931
 932         /**
 933          * @param string $code The code of the language for which to get the name
 934          * @param null|string $inLanguage Code of language in which to return the name
 935          *   (SELF::AS_AUTONYMS for autonyms)
 936          * @param string $include See fetchLanguageNames()
 937          * @return string Language name or empty
 938          * @since 1.20
 939          */
 940         public static function fetchLanguageName(
 941                 $code,
 942                 $inLanguage = self::AS_AUTONYMS,
 943                 $include = self::ALL
 944         ) {
 945                 $code = strtolower( $code );
 946                 $array = self::fetchLanguageNames( $inLanguage, $include );
 947                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 948         }
 949
 950         /**
 951          * Get a message from the MediaWiki namespace.
 952          *
 953          * @param string $msg Message name
 954          * @return string
 955          */
 956         public function getMessageFromDB( $msg ) {
 957                 return $this->msg( $msg )->text();
 958         }
 959
 960         /**
 961          * Get message object in this language. Only for use inside this class.
 962          *
 963          * @param string $msg Message name
 964          * @return Message
 965          */
 966         protected function msg( $msg ) {
 967                 return wfMessage( $msg )->inLanguage( $this );
 968         }
 969
 970         /**
 971          * @param string $key
 972          * @return string
 973          */
 974         public function getMonthName( $key ) {
 975                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 976         }
 977
 978         /**
 979          * @return array
 980          */
 981         public function getMonthNamesArray() {
 982                 $monthNames = [ '' ];
 983                 for ( $i = 1; $i < 13; $i++ ) {
 984                         $monthNames[] = $this->getMonthName( $i );
 985                 }
 986                 return $monthNames;
 987         }
 988
 989         /**
 990          * @param string $key
 991          * @return string
 992          */
 993         public function getMonthNameGen( $key ) {
 994                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 995         }
 996
 997         /**
 998          * @param string $key
 999          * @return string
1000          */
1001         public function getMonthAbbreviation( $key ) {
1002                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1003         }
1004
1005         /**
1006          * @return array
1007          */
1008         public function getMonthAbbreviationsArray() {
1009                 $monthNames = [ '' ];
1010                 for ( $i = 1; $i < 13; $i++ ) {
1011                         $monthNames[] = $this->getMonthAbbreviation( $i );
1012                 }
1013                 return $monthNames;
1014         }
1015
1016         /**
1017          * @param string $key
1018          * @return string
1019          */
1020         public function getWeekdayName( $key ) {
1021                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1022         }
1023
1024         /**
1025          * @param string $key
1026          * @return string
1027          */
1028         function getWeekdayAbbreviation( $key ) {
1029                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1030         }
1031
1032         /**
1033          * @param string $key
1034          * @return string
1035          */
1036         function getIranianCalendarMonthName( $key ) {
1037                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1038         }
1039
1040         /**
1041          * @param string $key
1042          * @return string
1043          */
1044         function getHebrewCalendarMonthName( $key ) {
1045                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1046         }
1047
1048         /**
1049          * @param string $key
1050          * @return string
1051          */
1052         function getHebrewCalendarMonthNameGen( $key ) {
1053                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1054         }
1055
1056         /**
1057          * @param string $key
1058          * @return string
1059          */
1060         function getHijriCalendarMonthName( $key ) {
1061                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1062         }
1063
1064         /**
1065          * Pass through result from $dateTimeObj->format()
1066          * @param DateTime|bool|null &$dateTimeObj
1067          * @param string $ts
1068          * @param DateTimeZone|bool|null $zone
1069          * @param string $code
1070          * @return string
1071          */
1072         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1073                 if ( !$dateTimeObj ) {
1074                         $dateTimeObj = DateTime::createFromFormat(
1075                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1076                         );
1077                 }
1078                 return $dateTimeObj->format( $code );
1079         }
1080
1081         /**
1082          * This is a workalike of PHP's date() function, but with better
1083          * internationalisation, a reduced set of format characters, and a better
1084          * escaping format.
1085          *
1086          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1087          * the PHP manual for definitions. There are a number of extensions, which
1088          * start with "x":
1089          *
1090          *    xn   Do not translate digits of the next numeric format character
1091          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1092          *    xr   Use roman numerals for the next numeric format character
1093          *    xh   Use hebrew numerals for the next numeric format character
1094          *    xx   Literal x
1095          *    xg   Genitive month name
1096          *
1097          *    xij  j (day number) in Iranian calendar
1098          *    xiF  F (month name) in Iranian calendar
1099          *    xin  n (month number) in Iranian calendar
1100          *    xiy  y (two digit year) in Iranian calendar
1101          *    xiY  Y (full year) in Iranian calendar
1102          *    xit  t (days in month) in Iranian calendar
1103          *    xiz  z (day of the year) in Iranian calendar
1104          *
1105          *    xjj  j (day number) in Hebrew calendar
1106          *    xjF  F (month name) in Hebrew calendar
1107          *    xjt  t (days in month) in Hebrew calendar
1108          *    xjx  xg (genitive month name) in Hebrew calendar
1109          *    xjn  n (month number) in Hebrew calendar
1110          *    xjY  Y (full year) in Hebrew calendar
1111          *
1112          *    xmj  j (day number) in Hijri calendar
1113          *    xmF  F (month name) in Hijri calendar
1114          *    xmn  n (month number) in Hijri calendar
1115          *    xmY  Y (full year) in Hijri calendar
1116          *
1117          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1118          *                       identical to the Gregorian calendar
1119          *    xoY  Y (full year) in Minguo calendar or Juche year.
1120          *                       Months and days are identical to the
1121          *                       Gregorian calendar
1122          *    xtY  Y (full year) in Japanese nengo. Months and days are
1123          *                       identical to the Gregorian calendar
1124          *
1125          * Characters enclosed in double quotes will be considered literal (with
1126          * the quotes themselves removed). Unmatched quotes will be considered
1127          * literal quotes. Example:
1128          *
1129          * "The month is" F       => The month is January
1130          * i's"                   => 20'11"
1131          *
1132          * Backslash escaping is also supported.
1133          *
1134          * Input timestamp is assumed to be pre-normalized to the desired local
1135          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1136          * $ts is UTC if $zone is not given.
1137          *
1138          * @param string $format
1139          * @param string $ts 14-character timestamp
1140          *      YYYYMMDDHHMMSS
1141          *      01234567890123
1142          * @param DateTimeZone|null $zone Timezone of $ts
1143          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1144          * Only makes sense if $ts is the current time.
1145          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1146          *
1147          * @throws MWException
1148          * @return string
1149          */
1150         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1151                 $s = '';
1152                 $raw = false;
1153                 $roman = false;
1154                 $hebrewNum = false;
1155                 $dateTimeObj = false;
1156                 $rawToggle = false;
1157                 $iranian = false;
1158                 $hebrew = false;
1159                 $hijri = false;
1160                 $thai = false;
1161                 $minguo = false;
1162                 $tenno = false;
1163
1164                 $usedSecond = false;
1165                 $usedMinute = false;
1166                 $usedHour = false;
1167                 $usedAMPM = false;
1168                 $usedDay = false;
1169                 $usedWeek = false;
1170                 $usedMonth = false;
1171                 $usedYear = false;
1172                 $usedISOYear = false;
1173                 $usedIsLeapYear = false;
1174
1175                 $usedHebrewMonth = false;
1176                 $usedIranianMonth = false;
1177                 $usedHijriMonth = false;
1178                 $usedHebrewYear = false;
1179                 $usedIranianYear = false;
1180                 $usedHijriYear = false;
1181                 $usedTennoYear = false;
1182
1183                 if ( strlen( $ts ) !== 14 ) {
1184                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1185                 }
1186
1187                 if ( !ctype_digit( $ts ) ) {
1188                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1189                 }
1190
1191                 $formatLength = strlen( $format );
1192                 for ( $p = 0; $p < $formatLength; $p++ ) {
1193                         $num = false;
1194                         $code = $format[$p];
1195                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1196                                 $code .= $format[++$p];
1197                         }
1198
1199                         if ( ( $code === 'xi'
1200                                         || $code === 'xj'
1201                                         || $code === 'xk'
1202                                         || $code === 'xm'
1203                                         || $code === 'xo'
1204                                         || $code === 'xt' )
1205                                 && $p < $formatLength - 1 ) {
1206                                 $code .= $format[++$p];
1207                         }
1208
1209                         switch ( $code ) {
1210                                 case 'xx':
1211                                         $s .= 'x';
1212                                         break;
1213                                 case 'xn':
1214                                         $raw = true;
1215                                         break;
1216                                 case 'xN':
1217                                         $rawToggle = !$rawToggle;
1218                                         break;
1219                                 case 'xr':
1220                                         $roman = true;
1221                                         break;
1222                                 case 'xh':
1223                                         $hebrewNum = true;
1224                                         break;
1225                                 case 'xg':
1226                                         $usedMonth = true;
1227                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1228                                         break;
1229                                 case 'xjx':
1230                                         $usedHebrewMonth = true;
1231                                         if ( !$hebrew ) {
1232                                                 $hebrew = self::tsToHebrew( $ts );
1233                                         }
1234                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1235                                         break;
1236                                 case 'd':
1237                                         $usedDay = true;
1238                                         $num = substr( $ts, 6, 2 );
1239                                         break;
1240                                 case 'D':
1241                                         $usedDay = true;
1242                                         $s .= $this->getWeekdayAbbreviation(
1243                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1244                                         );
1245                                         break;
1246                                 case 'j':
1247                                         $usedDay = true;
1248                                         $num = intval( substr( $ts, 6, 2 ) );
1249                                         break;
1250                                 case 'xij':
1251                                         $usedDay = true;
1252                                         if ( !$iranian ) {
1253                                                 $iranian = self::tsToIranian( $ts );
1254                                         }
1255                                         $num = $iranian[2];
1256                                         break;
1257                                 case 'xmj':
1258                                         $usedDay = true;
1259                                         if ( !$hijri ) {
1260                                                 $hijri = self::tsToHijri( $ts );
1261                                         }
1262                                         $num = $hijri[2];
1263                                         break;
1264                                 case 'xjj':
1265                                         $usedDay = true;
1266                                         if ( !$hebrew ) {
1267                                                 $hebrew = self::tsToHebrew( $ts );
1268                                         }
1269                                         $num = $hebrew[2];
1270                                         break;
1271                                 case 'l':
1272                                         $usedDay = true;
1273                                         $s .= $this->getWeekdayName(
1274                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1275                                         );
1276                                         break;
1277                                 case 'F':
1278                                         $usedMonth = true;
1279                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1280                                         break;
1281                                 case 'xiF':
1282                                         $usedIranianMonth = true;
1283                                         if ( !$iranian ) {
1284                                                 $iranian = self::tsToIranian( $ts );
1285                                         }
1286                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1287                                         break;
1288                                 case 'xmF':
1289                                         $usedHijriMonth = true;
1290                                         if ( !$hijri ) {
1291                                                 $hijri = self::tsToHijri( $ts );
1292                                         }
1293                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1294                                         break;
1295                                 case 'xjF':
1296                                         $usedHebrewMonth = true;
1297                                         if ( !$hebrew ) {
1298                                                 $hebrew = self::tsToHebrew( $ts );
1299                                         }
1300                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1301                                         break;
1302                                 case 'm':
1303                                         $usedMonth = true;
1304                                         $num = substr( $ts, 4, 2 );
1305                                         break;
1306                                 case 'M':
1307                                         $usedMonth = true;
1308                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1309                                         break;
1310                                 case 'n':
1311                                         $usedMonth = true;
1312                                         $num = intval( substr( $ts, 4, 2 ) );
1313                                         break;
1314                                 case 'xin':
1315                                         $usedIranianMonth = true;
1316                                         if ( !$iranian ) {
1317                                                 $iranian = self::tsToIranian( $ts );
1318                                         }
1319                                         $num = $iranian[1];
1320                                         break;
1321                                 case 'xmn':
1322                                         $usedHijriMonth = true;
1323                                         if ( !$hijri ) {
1324                                                 $hijri = self::tsToHijri( $ts );
1325                                         }
1326                                         $num = $hijri[1];
1327                                         break;
1328                                 case 'xjn':
1329                                         $usedHebrewMonth = true;
1330                                         if ( !$hebrew ) {
1331                                                 $hebrew = self::tsToHebrew( $ts );
1332                                         }
1333                                         $num = $hebrew[1];
1334                                         break;
1335                                 case 'xjt':
1336                                         $usedHebrewMonth = true;
1337                                         if ( !$hebrew ) {
1338                                                 $hebrew = self::tsToHebrew( $ts );
1339                                         }
1340                                         $num = $hebrew[3];
1341                                         break;
1342                                 case 'Y':
1343                                         $usedYear = true;
1344                                         $num = substr( $ts, 0, 4 );
1345                                         break;
1346                                 case 'xiY':
1347                                         $usedIranianYear = true;
1348                                         if ( !$iranian ) {
1349                                                 $iranian = self::tsToIranian( $ts );
1350                                         }
1351                                         $num = $iranian[0];
1352                                         break;
1353                                 case 'xmY':
1354                                         $usedHijriYear = true;
1355                                         if ( !$hijri ) {
1356                                                 $hijri = self::tsToHijri( $ts );
1357                                         }
1358                                         $num = $hijri[0];
1359                                         break;
1360                                 case 'xjY':
1361                                         $usedHebrewYear = true;
1362                                         if ( !$hebrew ) {
1363                                                 $hebrew = self::tsToHebrew( $ts );
1364                                         }
1365                                         $num = $hebrew[0];
1366                                         break;
1367                                 case 'xkY':
1368                                         $usedYear = true;
1369                                         if ( !$thai ) {
1370                                                 $thai = self::tsToYear( $ts, 'thai' );
1371                                         }
1372                                         $num = $thai[0];
1373                                         break;
1374                                 case 'xoY':
1375                                         $usedYear = true;
1376                                         if ( !$minguo ) {
1377                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1378                                         }
1379                                         $num = $minguo[0];
1380                                         break;
1381                                 case 'xtY':
1382                                         $usedTennoYear = true;
1383                                         if ( !$tenno ) {
1384                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1385                                         }
1386                                         $num = $tenno[0];
1387                                         break;
1388                                 case 'y':
1389                                         $usedYear = true;
1390                                         $num = substr( $ts, 2, 2 );
1391                                         break;
1392                                 case 'xiy':
1393                                         $usedIranianYear = true;
1394                                         if ( !$iranian ) {
1395                                                 $iranian = self::tsToIranian( $ts );
1396                                         }
1397                                         $num = substr( $iranian[0], -2 );
1398                                         break;
1399                                 case 'xit':
1400                                         $usedIranianYear = true;
1401                                         if ( !$iranian ) {
1402                                                 $iranian = self::tsToIranian( $ts );
1403                                         }
1404                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1405                                         break;
1406                                 case 'xiz':
1407                                         $usedIranianYear = true;
1408                                         if ( !$iranian ) {
1409                                                 $iranian = self::tsToIranian( $ts );
1410                                         }
1411                                         $num = $iranian[3];
1412                                         break;
1413                                 case 'a':
1414                                         $usedAMPM = true;
1415                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1416                                         break;
1417                                 case 'A':
1418                                         $usedAMPM = true;
1419                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1420                                         break;
1421                                 case 'g':
1422                                         $usedHour = true;
1423                                         $h = substr( $ts, 8, 2 );
1424                                         $num = $h % 12 ?: 12;
1425                                         break;
1426                                 case 'G':
1427                                         $usedHour = true;
1428                                         $num = intval( substr( $ts, 8, 2 ) );
1429                                         break;
1430                                 case 'h':
1431                                         $usedHour = true;
1432                                         $h = substr( $ts, 8, 2 );
1433                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1434                                         break;
1435                                 case 'H':
1436                                         $usedHour = true;
1437                                         $num = substr( $ts, 8, 2 );
1438                                         break;
1439                                 case 'i':
1440                                         $usedMinute = true;
1441                                         $num = substr( $ts, 10, 2 );
1442                                         break;
1443                                 case 's':
1444                                         $usedSecond = true;
1445                                         $num = substr( $ts, 12, 2 );
1446                                         break;
1447                                 case 'c':
1448                                 case 'r':
1449                                         $usedSecond = true;
1450                                         // fall through
1451                                 case 'e':
1452                                 case 'O':
1453                                 case 'P':
1454                                 case 'T':
1455                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1456                                         break;
1457                                 case 'w':
1458                                 case 'N':
1459                                 case 'z':
1460                                         $usedDay = true;
1461                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1462                                         break;
1463                                 case 'W':
1464                                         $usedWeek = true;
1465                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1466                                         break;
1467                                 case 't':
1468                                         $usedMonth = true;
1469                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1470                                         break;
1471                                 case 'L':
1472                                         $usedIsLeapYear = true;
1473                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1474                                         break;
1475                                 case 'o':
1476                                         $usedISOYear = true;
1477                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1478                                         break;
1479                                 case 'U':
1480                                         $usedSecond = true;
1481                                         // fall through
1482                                 case 'I':
1483                                 case 'Z':
1484                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1485                                         break;
1486                                 case '\\':
1487                                         # Backslash escaping
1488                                         if ( $p < $formatLength - 1 ) {
1489                                                 $s .= $format[++$p];
1490                                         } else {
1491                                                 $s .= '\\';
1492                                         }
1493                                         break;
1494                                 case '"':
1495                                         # Quoted literal
1496                                         if ( $p < $formatLength - 1 ) {
1497                                                 $endQuote = strpos( $format, '"', $p + 1 );
1498                                                 if ( $endQuote === false ) {
1499                                                         # No terminating quote, assume literal "
1500                                                         $s .= '"';
1501                                                 } else {
1502                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1503                                                         $p = $endQuote;
1504                                                 }
1505                                         } else {
1506                                                 # Quote at end of string, assume literal "
1507                                                 $s .= '"';
1508                                         }
1509                                         break;
1510                                 default:
1511                                         $s .= $format[$p];
1512                         }
1513                         if ( $num !== false ) {
1514                                 if ( $rawToggle || $raw ) {
1515                                         $s .= $num;
1516                                         $raw = false;
1517                                 } elseif ( $roman ) {
1518                                         $s .= self::romanNumeral( $num );
1519                                         $roman = false;
1520                                 } elseif ( $hebrewNum ) {
1521                                         $s .= self::hebrewNumeral( $num );
1522                                         $hebrewNum = false;
1523                                 } else {
1524                                         $s .= $this->formatNum( $num, true );
1525                                 }
1526                         }
1527                 }
1528
1529                 if ( $ttl === 'unused' ) {
1530                         // No need to calculate the TTL, the caller wont use it anyway.
1531                 } elseif ( $usedSecond ) {
1532                         $ttl = 1;
1533                 } elseif ( $usedMinute ) {
1534                         $ttl = 60 - substr( $ts, 12, 2 );
1535                 } elseif ( $usedHour ) {
1536                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1537                 } elseif ( $usedAMPM ) {
1538                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1539                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1540                 } elseif (
1541                         $usedDay ||
1542                         $usedHebrewMonth ||
1543                         $usedIranianMonth ||
1544                         $usedHijriMonth ||
1545                         $usedHebrewYear ||
1546                         $usedIranianYear ||
1547                         $usedHijriYear ||
1548                         $usedTennoYear
1549                 ) {
1550                         // @todo Someone who understands the non-Gregorian calendars
1551                         // should write proper logic for them so that they don't need purged every day.
1552                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1553                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1554                 } else {
1555                         $possibleTtls = [];
1556                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1557                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1558                         if ( $usedWeek ) {
1559                                 $possibleTtls[] =
1560                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1561                                         $timeRemainingInDay;
1562                         } elseif ( $usedISOYear ) {
1563                                 // December 28th falls on the last ISO week of the year, every year.
1564                                 // The last ISO week of a year can be 52 or 53.
1565                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1566                                         'Ymd',
1567                                         substr( $ts, 0, 4 ) . '1228',
1568                                         $zone ?: new DateTimeZone( 'UTC' )
1569                                 )->format( 'W' );
1570                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1571                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1572                                 $timeRemainingInWeek =
1573                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1574                                         + $timeRemainingInDay;
1575                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1576                         }
1577
1578                         if ( $usedMonth ) {
1579                                 $possibleTtls[] =
1580                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1581                                                 substr( $ts, 6, 2 ) ) * 86400
1582                                         + $timeRemainingInDay;
1583                         } elseif ( $usedYear ) {
1584                                 $possibleTtls[] =
1585                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1586                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1587                                         + $timeRemainingInDay;
1588                         } elseif ( $usedIsLeapYear ) {
1589                                 $year = substr( $ts, 0, 4 );
1590                                 $timeRemainingInYear =
1591                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1592                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1593                                         + $timeRemainingInDay;
1594                                 $mod = $year % 4;
1595                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1596                                         // this isn't a leap year. see when the next one starts
1597                                         $nextCandidate = $year - $mod + 4;
1598                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1599                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1600                                                         $timeRemainingInYear;
1601                                         } else {
1602                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1603                                                         $timeRemainingInYear;
1604                                         }
1605                                 } else {
1606                                         // this is a leap year, so the next year isn't
1607                                         $possibleTtls[] = $timeRemainingInYear;
1608                                 }
1609                         }
1610
1611                         if ( $possibleTtls ) {
1612                                 $ttl = min( $possibleTtls );
1613                         }
1614                 }
1615
1616                 return $s;
1617         }
1618
1619         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1620         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1621
1622         /**
1623          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1624          * Gregorian dates to Iranian dates. Originally written in C, it
1625          * is released under the terms of GNU Lesser General Public
1626          * License. Conversion to PHP was performed by Niklas Laxström.
1627          *
1628          * Link: http://www.farsiweb.info/jalali/jalali.c
1629          *
1630          * @param string $ts
1631          *
1632          * @return int[]
1633          */
1634         private static function tsToIranian( $ts ) {
1635                 $gy = substr( $ts, 0, 4 ) - 1600;
1636                 $gm = substr( $ts, 4, 2 ) - 1;
1637                 $gd = substr( $ts, 6, 2 ) - 1;
1638
1639                 # Days passed from the beginning (including leap years)
1640                 $gDayNo = 365 * $gy
1641                         + floor( ( $gy + 3 ) / 4 )
1642                         - floor( ( $gy + 99 ) / 100 )
1643                         + floor( ( $gy + 399 ) / 400 );
1644
1645                 // Add days of the past months of this year
1646                 for ( $i = 0; $i < $gm; $i++ ) {
1647                         $gDayNo += self::$GREG_DAYS[$i];
1648                 }
1649
1650                 // Leap years
1651                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1652                         $gDayNo++;
1653                 }
1654
1655                 // Days passed in current month
1656                 $gDayNo += (int)$gd;
1657
1658                 $jDayNo = $gDayNo - 79;
1659
1660                 $jNp = floor( $jDayNo / 12053 );
1661                 $jDayNo %= 12053;
1662
1663                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1664                 $jDayNo %= 1461;
1665
1666                 if ( $jDayNo >= 366 ) {
1667                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1668                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1669                 }
1670
1671                 $jz = $jDayNo;
1672
1673                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1674                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1675                 }
1676
1677                 $jm = $i + 1;
1678                 $jd = $jDayNo + 1;
1679
1680                 return [ $jy, $jm, $jd, $jz ];
1681         }
1682
1683         /**
1684          * Converting Gregorian dates to Hijri dates.
1685          *
1686          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1687          *
1688          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1689          *
1690          * @param string $ts
1691          *
1692          * @return int[]
1693          */
1694         private static function tsToHijri( $ts ) {
1695                 $year = substr( $ts, 0, 4 );
1696                 $month = substr( $ts, 4, 2 );
1697                 $day = substr( $ts, 6, 2 );
1698
1699                 $zyr = $year;
1700                 $zd = $day;
1701                 $zm = $month;
1702                 $zy = $zyr;
1703
1704                 if (
1705                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1706                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1707                 ) {
1708                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1709                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1710                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1711                                         $zd - 32075;
1712                 } else {
1713                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1714                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1715                 }
1716
1717                 $zl = $zjd - 1948440 + 10632;
1718                 $zn = (int)( ( $zl - 1 ) / 10631 );
1719                 $zl = $zl - 10631 * $zn + 354;
1720                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1721                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1722                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1723                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1724                 $zm = (int)( ( 24 * $zl ) / 709 );
1725                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1726                 $zy = 30 * $zn + $zj - 30;
1727
1728                 return [ $zy, $zm, $zd ];
1729         }
1730
1731         /**
1732          * Converting Gregorian dates to Hebrew dates.
1733          *
1734          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1735          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1736          * to translate the relevant functions into PHP and release them under
1737          * GNU GPL.
1738          *
1739          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1740          * and Adar II is 14. In a non-leap year, Adar is 6.
1741          *
1742          * @param string $ts
1743          *
1744          * @return int[]
1745          */
1746         private static function tsToHebrew( $ts ) {
1747                 # Parse date
1748                 $year = substr( $ts, 0, 4 );
1749                 $month = substr( $ts, 4, 2 );
1750                 $day = substr( $ts, 6, 2 );
1751
1752                 # Calculate Hebrew year
1753                 $hebrewYear = $year + 3760;
1754
1755                 # Month number when September = 1, August = 12
1756                 $month += 4;
1757                 if ( $month > 12 ) {
1758                         # Next year
1759                         $month -= 12;
1760                         $year++;
1761                         $hebrewYear++;
1762                 }
1763
1764                 # Calculate day of year from 1 September
1765                 $dayOfYear = $day;
1766                 for ( $i = 1; $i < $month; $i++ ) {
1767                         if ( $i == 6 ) {
1768                                 # February
1769                                 $dayOfYear += 28;
1770                                 # Check if the year is leap
1771                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1772                                         $dayOfYear++;
1773                                 }
1774                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1775                                 $dayOfYear += 30;
1776                         } else {
1777                                 $dayOfYear += 31;
1778                         }
1779                 }
1780
1781                 # Calculate the start of the Hebrew year
1782                 $start = self::hebrewYearStart( $hebrewYear );
1783
1784                 # Calculate next year's start
1785                 if ( $dayOfYear <= $start ) {
1786                         # Day is before the start of the year - it is the previous year
1787                         # Next year's start
1788                         $nextStart = $start;
1789                         # Previous year
1790                         $year--;
1791                         $hebrewYear--;
1792                         # Add days since previous year's 1 September
1793                         $dayOfYear += 365;
1794                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1795                                 # Leap year
1796                                 $dayOfYear++;
1797                         }
1798                         # Start of the new (previous) year
1799                         $start = self::hebrewYearStart( $hebrewYear );
1800                 } else {
1801                         # Next year's start
1802                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1803                 }
1804
1805                 # Calculate Hebrew day of year
1806                 $hebrewDayOfYear = $dayOfYear - $start;
1807
1808                 # Difference between year's days
1809                 $diff = $nextStart - $start;
1810                 # Add 12 (or 13 for leap years) days to ignore the difference between
1811                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1812                 # difference is only about the year type
1813                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1814                         $diff += 13;
1815                 } else {
1816                         $diff += 12;
1817                 }
1818
1819                 # Check the year pattern, and is leap year
1820                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1821                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1822                 # and non-leap years
1823                 $yearPattern = $diff % 30;
1824                 # Check if leap year
1825                 $isLeap = $diff >= 30;
1826
1827                 # Calculate day in the month from number of day in the Hebrew year
1828                 # Don't check Adar - if the day is not in Adar, we will stop before;
1829                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1830                 $hebrewDay = $hebrewDayOfYear;
1831                 $hebrewMonth = 1;
1832                 $days = 0;
1833                 while ( $hebrewMonth <= 12 ) {
1834                         # Calculate days in this month
1835                         if ( $isLeap && $hebrewMonth == 6 ) {
1836                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1837                                 $days = 30;
1838                                 if ( $hebrewDay <= $days ) {
1839                                         # Day in Adar I
1840                                         $hebrewMonth = 13;
1841                                 } else {
1842                                         # Subtract the days of Adar I
1843                                         $hebrewDay -= $days;
1844                                         # Try Adar II
1845                                         $days = 29;
1846                                         if ( $hebrewDay <= $days ) {
1847                                                 # Day in Adar II
1848                                                 $hebrewMonth = 14;
1849                                         }
1850                                 }
1851                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1852                                 # Cheshvan in a complete year (otherwise as the rule below)
1853                                 $days = 30;
1854                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1855                                 # Kislev in an incomplete year (otherwise as the rule below)
1856                                 $days = 29;
1857                         } else {
1858                                 # Odd months have 30 days, even have 29
1859                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1860                         }
1861                         if ( $hebrewDay <= $days ) {
1862                                 # In the current month
1863                                 break;
1864                         } else {
1865                                 # Subtract the days of the current month
1866                                 $hebrewDay -= $days;
1867                                 # Try in the next month
1868                                 $hebrewMonth++;
1869                         }
1870                 }
1871
1872                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1873         }
1874
1875         /**
1876          * This calculates the Hebrew year start, as days since 1 September.
1877          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1878          * Used for Hebrew date.
1879          *
1880          * @param int $year
1881          *
1882          * @return string
1883          */
1884         private static function hebrewYearStart( $year ) {
1885                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1886                 $b = intval( ( $year - 1 ) % 4 );
1887                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1888                 if ( $m < 0 ) {
1889                         $m--;
1890                 }
1891                 $Mar = intval( $m );
1892                 if ( $m < 0 ) {
1893                         $m++;
1894                 }
1895                 $m -= $Mar;
1896
1897                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1898                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1899                         $Mar++;
1900                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1901                         $Mar += 2;
1902                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1903                         $Mar++;
1904                 }
1905
1906                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1907                 return $Mar;
1908         }
1909
1910         /**
1911          * Algorithm to convert Gregorian dates to Thai solar dates,
1912          * Minguo dates or Minguo dates.
1913          *
1914          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1915          *       https://en.wikipedia.org/wiki/Minguo_calendar
1916          *       https://en.wikipedia.org/wiki/Japanese_era_name
1917          *
1918          * @param string $ts 14-character timestamp
1919          * @param string $cName Calender name
1920          * @return array Converted year, month, day
1921          */
1922         private static function tsToYear( $ts, $cName ) {
1923                 $gy = substr( $ts, 0, 4 );
1924                 $gm = substr( $ts, 4, 2 );
1925                 $gd = substr( $ts, 6, 2 );
1926
1927                 if ( !strcmp( $cName, 'thai' ) ) {
1928                         # Thai solar dates
1929                         # Add 543 years to the Gregorian calendar
1930                         # Months and days are identical
1931                         $gy_offset = $gy + 543;
1932                         # fix for dates between 1912 and 1941
1933                         # https://en.wikipedia.org/?oldid=836596673#New_year
1934                         if ( $gy >= 1912 && $gy <= 1940 ) {
1935                                 if ( $gm <= 3 ) {
1936                                         $gy_offset--;
1937                                 }
1938                                 $gm = ( $gm - 3 ) % 12;
1939                         }
1940                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1941                         # Minguo dates
1942                         # Deduct 1911 years from the Gregorian calendar
1943                         # Months and days are identical
1944                         $gy_offset = $gy - 1911;
1945                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1946                         # Nengō dates up to Meiji period
1947                         # Deduct years from the Gregorian calendar
1948                         # depending on the nengo periods
1949                         # Months and days are identical
1950                         if ( ( $gy < 1912 )
1951                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1952                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1953                         ) {
1954                                 # Meiji period
1955                                 $gy_gannen = $gy - 1868 + 1;
1956                                 $gy_offset = $gy_gannen;
1957                                 if ( $gy_gannen == 1 ) {
1958                                         $gy_offset = '元';
1959                                 }
1960                                 $gy_offset = '明治' . $gy_offset;
1961                         } elseif (
1962                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1963                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1964                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1965                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1966                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1967                         ) {
1968                                 # Taishō period
1969                                 $gy_gannen = $gy - 1912 + 1;
1970                                 $gy_offset = $gy_gannen;
1971                                 if ( $gy_gannen == 1 ) {
1972                                         $gy_offset = '元';
1973                                 }
1974                                 $gy_offset = '大正' . $gy_offset;
1975                         } elseif (
1976                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1977                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1978                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1979                         ) {
1980                                 # Shōwa period
1981                                 $gy_gannen = $gy - 1926 + 1;
1982                                 $gy_offset = $gy_gannen;
1983                                 if ( $gy_gannen == 1 ) {
1984                                         $gy_offset = '元';
1985                                 }
1986                                 $gy_offset = '昭和' . $gy_offset;
1987                         } else {
1988                                 # Heisei period
1989                                 $gy_gannen = $gy - 1989 + 1;
1990                                 $gy_offset = $gy_gannen;
1991                                 if ( $gy_gannen == 1 ) {
1992                                         $gy_offset = '元';
1993                                 }
1994                                 $gy_offset = '平成' . $gy_offset;
1995                         }
1996                 } else {
1997                         $gy_offset = $gy;
1998                 }
1999
2000                 return [ $gy_offset, $gm, $gd ];
2001         }
2002
2003         /**
2004          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2005          *
2006          * This is the rule the BIDI algorithm uses to determine the directionality of
2007          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2008          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2009          *
2010          * TODO: Does not handle BIDI control characters inside the text.
2011          * TODO: Does not handle unallocated characters.
2012          *
2013          * @param string $text Text to test
2014          * @return null|string Directionality ('ltr' or 'rtl') or null
2015          */
2016         private static function strongDirFromContent( $text = '' ) {
2017                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2018                         return null;
2019                 }
2020                 if ( $matches[1] === '' ) {
2021                         return 'rtl';
2022                 }
2023                 return 'ltr';
2024         }
2025
2026         /**
2027          * Roman number formatting up to 10000
2028          *
2029          * @param int $num
2030          *
2031          * @return string
2032          */
2033         static function romanNumeral( $num ) {
2034                 static $table = [
2035                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2036                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2037                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2038                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2039                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2040                 ];
2041
2042                 $num = intval( $num );
2043                 if ( $num > 10000 || $num <= 0 ) {
2044                         return $num;
2045                 }
2046
2047                 $s = '';
2048                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2049                         if ( $num >= $pow10 ) {
2050                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2051                         }
2052                         $num = $num % $pow10;
2053                 }
2054                 return $s;
2055         }
2056
2057         /**
2058          * Hebrew Gematria number formatting up to 9999
2059          *
2060          * @param int $num
2061          *
2062          * @return string
2063          */
2064         static function hebrewNumeral( $num ) {
2065                 static $table = [
2066                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2067                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2068                         [ '',
2069                                 [ 'ק' ],
2070                                 [ 'ר' ],
2071                                 [ 'ש' ],
2072                                 [ 'ת' ],
2073                                 [ 'ת', 'ק' ],
2074                                 [ 'ת', 'ר' ],
2075                                 [ 'ת', 'ש' ],
2076                                 [ 'ת', 'ת' ],
2077                                 [ 'ת', 'ת', 'ק' ],
2078                                 [ 'ת', 'ת', 'ר' ],
2079                         ],
2080                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2081                 ];
2082
2083                 $num = intval( $num );
2084                 if ( $num > 9999 || $num <= 0 ) {
2085                         return $num;
2086                 }
2087
2088                 // Round thousands have special notations
2089                 if ( $num === 1000 ) {
2090                         return "א' אלף";
2091                 } elseif ( $num % 1000 === 0 ) {
2092                         return $table[0][$num / 1000] . "' אלפים";
2093                 }
2094
2095                 $letters = [];
2096
2097                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2098                         if ( $num >= $pow10 ) {
2099                                 if ( $num === 15 || $num === 16 ) {
2100                                         $letters[] = $table[0][9];
2101                                         $letters[] = $table[0][$num - 9];
2102                                         $num = 0;
2103                                 } else {
2104                                         $letters = array_merge(
2105                                                 $letters,
2106                                                 (array)$table[$i][intval( $num / $pow10 )]
2107                                         );
2108
2109                                         if ( $pow10 === 1000 ) {
2110                                                 $letters[] = "'";
2111                                         }
2112                                 }
2113                         }
2114
2115                         $num = $num % $pow10;
2116                 }
2117
2118                 $preTransformLength = count( $letters );
2119                 if ( $preTransformLength === 1 ) {
2120                         // Add geresh (single quote) to one-letter numbers
2121                         $letters[] = "'";
2122                 } else {
2123                         $lastIndex = $preTransformLength - 1;
2124                         $letters[$lastIndex] = str_replace(
2125                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2126                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2127                                 $letters[$lastIndex]
2128                         );
2129
2130                         // Add gershayim (double quote) to multiple-letter numbers,
2131                         // but exclude numbers with only one letter after the thousands
2132                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2133                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2134                                 $letters[] = "'";
2135                         } else {
2136                                 array_splice( $letters, -1, 0, '"' );
2137                         }
2138                 }
2139
2140                 return implode( $letters );
2141         }
2142
2143         /**
2144          * Used by date() and time() to adjust the time output.
2145          *
2146          * @param string $ts The time in date('YmdHis') format
2147          * @param mixed $tz Adjust the time by this amount (default false, mean we
2148          *   get user timecorrection setting)
2149          * @return int
2150          */
2151         public function userAdjust( $ts, $tz = false ) {
2152                 global $wgUser, $wgLocalTZoffset;
2153
2154                 if ( $tz === false ) {
2155                         $tz = $wgUser->getOption( 'timecorrection' );
2156                 }
2157
2158                 $data = explode( '|', $tz, 3 );
2159
2160                 if ( $data[0] == 'ZoneInfo' ) {
2161                         try {
2162                                 $userTZ = new DateTimeZone( $data[2] );
2163                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2164                                 $date->setTimezone( $userTZ );
2165                                 return $date->format( 'YmdHis' );
2166                         } catch ( Exception $e ) {
2167                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2168                                 $data[0] = 'Offset';
2169                         }
2170                 }
2171
2172                 if ( $data[0] == 'System' || $tz == '' ) {
2173                         # Global offset in minutes.
2174                         $minDiff = $wgLocalTZoffset;
2175                 } elseif ( $data[0] == 'Offset' ) {
2176                         $minDiff = intval( $data[1] );
2177                 } else {
2178                         $data = explode( ':', $tz );
2179                         if ( count( $data ) == 2 ) {
2180                                 $data[0] = intval( $data[0] );
2181                                 $data[1] = intval( $data[1] );
2182                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2183                                 if ( $data[0] < 0 ) {
2184                                         $minDiff = -$minDiff;
2185                                 }
2186                         } else {
2187                                 $minDiff = intval( $data[0] ) * 60;
2188                         }
2189                 }
2190
2191                 # No difference ? Return time unchanged
2192                 if ( $minDiff == 0 ) {
2193                         return $ts;
2194                 }
2195
2196                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2197                 # Generate an adjusted date; take advantage of the fact that mktime
2198                 # will normalize out-of-range values so we don't have to split $minDiff
2199                 # into hours and minutes.
2200                 $t = mktime( (
2201                         (int)substr( $ts, 8, 2 ) ), # Hours
2202                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2203                         (int)substr( $ts, 12, 2 ), # Seconds
2204                         (int)substr( $ts, 4, 2 ), # Month
2205                         (int)substr( $ts, 6, 2 ), # Day
2206                         (int)substr( $ts, 0, 4 ) ); # Year
2207
2208                 $date = date( 'YmdHis', $t );
2209                 Wikimedia\restoreWarnings();
2210
2211                 return $date;
2212         }
2213
2214         /**
2215          * This is meant to be used by time(), date(), and timeanddate() to get
2216          * the date preference they're supposed to use, it should be used in
2217          * all children.
2218          *
2219          *     function timeanddate([...], $format = true) {
2220          *       $datePreference = $this->dateFormat($format);
2221          *       [...]
2222          *     }
2223          *
2224          * @param int|string|bool $usePrefs If true, the user's preference is used
2225          *   if false, the site/language default is used
2226          *   if int/string, assumed to be a format.
2227          * @return string
2228          */
2229         function dateFormat( $usePrefs = true ) {
2230                 global $wgUser;
2231
2232                 if ( is_bool( $usePrefs ) ) {
2233                         if ( $usePrefs ) {
2234                                 $datePreference = $wgUser->getDatePreference();
2235                         } else {
2236                                 $datePreference = (string)User::getDefaultOption( 'date' );
2237                         }
2238                 } else {
2239                         $datePreference = (string)$usePrefs;
2240                 }
2241
2242                 // return int
2243                 if ( $datePreference == '' ) {
2244                         return 'default';
2245                 }
2246
2247                 return $datePreference;
2248         }
2249
2250         /**
2251          * Get a format string for a given type and preference
2252          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2253          * @param string $pref The format name as it appears in Messages*.php under
2254          *  $datePreferences.
2255          *
2256          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2257          *
2258          * @return string
2259          */
2260         function getDateFormatString( $type, $pref ) {
2261                 $wasDefault = false;
2262                 if ( $pref == 'default' ) {
2263                         $wasDefault = true;
2264                         $pref = $this->getDefaultDateFormat();
2265                 }
2266
2267                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2268                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2269
2270                         if ( $type === 'pretty' && $df === null ) {
2271                                 $df = $this->getDateFormatString( 'date', $pref );
2272                         }
2273
2274                         if ( !$wasDefault && $df === null ) {
2275                                 $pref = $this->getDefaultDateFormat();
2276                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2277                         }
2278
2279                         $this->dateFormatStrings[$type][$pref] = $df;
2280                 }
2281                 return $this->dateFormatStrings[$type][$pref];
2282         }
2283
2284         /**
2285          * @param string $ts The time format which needs to be turned into a
2286          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2287          * @param bool $adj Whether to adjust the time output according to the
2288          *   user configured offset ($timecorrection)
2289          * @param mixed $format True to use user's date format preference
2290          * @param string|bool $timecorrection The time offset as returned by
2291          *   validateTimeZone() in Special:Preferences
2292          * @return string
2293          */
2294         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2295                 $ts = wfTimestamp( TS_MW, $ts );
2296                 if ( $adj ) {
2297                         $ts = $this->userAdjust( $ts, $timecorrection );
2298                 }
2299                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2300                 return $this->sprintfDate( $df, $ts );
2301         }
2302
2303         /**
2304          * @param string $ts The time format which needs to be turned into a
2305          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2306          * @param bool $adj Whether to adjust the time output according to the
2307          *   user configured offset ($timecorrection)
2308          * @param mixed $format True to use user's date format preference
2309          * @param string|bool $timecorrection The time offset as returned by
2310          *   validateTimeZone() in Special:Preferences
2311          * @return string
2312          */
2313         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2314                 $ts = wfTimestamp( TS_MW, $ts );
2315                 if ( $adj ) {
2316                         $ts = $this->userAdjust( $ts, $timecorrection );
2317                 }
2318                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2319                 return $this->sprintfDate( $df, $ts );
2320         }
2321
2322         /**
2323          * @param string $ts The time format which needs to be turned into a
2324          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2325          * @param bool $adj Whether to adjust the time output according to the
2326          *   user configured offset ($timecorrection)
2327          * @param mixed $format What format to return, if it's false output the
2328          *   default one (default true)
2329          * @param string|bool $timecorrection The time offset as returned by
2330          *   validateTimeZone() in Special:Preferences
2331          * @return string
2332          */
2333         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2334                 $ts = wfTimestamp( TS_MW, $ts );
2335                 if ( $adj ) {
2336                         $ts = $this->userAdjust( $ts, $timecorrection );
2337                 }
2338                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2339                 return $this->sprintfDate( $df, $ts );
2340         }
2341
2342         /**
2343          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2344          *
2345          * @since 1.20
2346          *
2347          * @param int $seconds The amount of seconds.
2348          * @param array $chosenIntervals The intervals to enable.
2349          *
2350          * @return string
2351          */
2352         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2353                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2354
2355                 $segments = [];
2356
2357                 foreach ( $intervals as $intervalName => $intervalValue ) {
2358                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2359                         // duration-years, duration-decades, duration-centuries, duration-millennia
2360                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2361                         $segments[] = $message->inLanguage( $this )->escaped();
2362                 }
2363
2364                 return $this->listToText( $segments );
2365         }
2366
2367         /**
2368          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2369          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2370          *
2371          * @since 1.20
2372          *
2373          * @param int $seconds The amount of seconds.
2374          * @param array $chosenIntervals The intervals to enable.
2375          *
2376          * @return array
2377          */
2378         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2379                 if ( empty( $chosenIntervals ) ) {
2380                         $chosenIntervals = [
2381                                 'millennia',
2382                                 'centuries',
2383                                 'decades',
2384                                 'years',
2385                                 'days',
2386                                 'hours',
2387                                 'minutes',
2388                                 'seconds'
2389                         ];
2390                 }
2391
2392                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2393                 $sortedNames = array_keys( $intervals );
2394                 $smallestInterval = array_pop( $sortedNames );
2395
2396                 $segments = [];
2397
2398                 foreach ( $intervals as $name => $length ) {
2399                         $value = floor( $seconds / $length );
2400
2401                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2402                                 $seconds -= $value * $length;
2403                                 $segments[$name] = $value;
2404                         }
2405                 }
2406
2407                 return $segments;
2408         }
2409
2410         /**
2411          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2412          *
2413          * @param string $type Can be 'date', 'time' or 'both'
2414          * @param string $ts The time format which needs to be turned into a
2415          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2416          * @param User $user User object used to get preferences for timezone and format
2417          * @param array $options Array, can contain the following keys:
2418          *   - 'timecorrection': time correction, can have the following values:
2419          *     - true: use user's preference
2420          *     - false: don't use time correction
2421          *     - int: value of time correction in minutes
2422          *   - 'format': format to use, can have the following values:
2423          *     - true: use user's preference
2424          *     - false: use default preference
2425          *     - string: format to use
2426          * @since 1.19
2427          * @return string
2428          */
2429         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2430                 $ts = wfTimestamp( TS_MW, $ts );
2431                 $options += [ 'timecorrection' => true, 'format' => true ];
2432                 if ( $options['timecorrection'] !== false ) {
2433                         if ( $options['timecorrection'] === true ) {
2434                                 $offset = $user->getOption( 'timecorrection' );
2435                         } else {
2436                                 $offset = $options['timecorrection'];
2437                         }
2438                         $ts = $this->userAdjust( $ts, $offset );
2439                 }
2440                 if ( $options['format'] === true ) {
2441                         $format = $user->getDatePreference();
2442                 } else {
2443                         $format = $options['format'];
2444                 }
2445                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2446                 return $this->sprintfDate( $df, $ts );
2447         }
2448
2449         /**
2450          * Get the formatted date for the given timestamp and formatted for
2451          * the given user.
2452          *
2453          * @param mixed $ts Mixed: the time format which needs to be turned into a
2454          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2455          * @param User $user User object used to get preferences for timezone and format
2456          * @param array $options Array, can contain the following keys:
2457          *   - 'timecorrection': time correction, can have the following values:
2458          *     - true: use user's preference
2459          *     - false: don't use time correction
2460          *     - int: value of time correction in minutes
2461          *   - 'format': format to use, can have the following values:
2462          *     - true: use user's preference
2463          *     - false: use default preference
2464          *     - string: format to use
2465          * @since 1.19
2466          * @return string
2467          */
2468         public function userDate( $ts, User $user, array $options = [] ) {
2469                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2470         }
2471
2472         /**
2473          * Get the formatted time for the given timestamp and formatted for
2474          * the given user.
2475          *
2476          * @param mixed $ts The time format which needs to be turned into a
2477          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2478          * @param User $user User object used to get preferences for timezone and format
2479          * @param array $options Array, can contain the following keys:
2480          *   - 'timecorrection': time correction, can have the following values:
2481          *     - true: use user's preference
2482          *     - false: don't use time correction
2483          *     - int: value of time correction in minutes
2484          *   - 'format': format to use, can have the following values:
2485          *     - true: use user's preference
2486          *     - false: use default preference
2487          *     - string: format to use
2488          * @since 1.19
2489          * @return string
2490          */
2491         public function userTime( $ts, User $user, array $options = [] ) {
2492                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2493         }
2494
2495         /**
2496          * Get the formatted date and time for the given timestamp and formatted for
2497          * the given user.
2498          *
2499          * @param mixed $ts The time format which needs to be turned into a
2500          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2501          * @param User $user User object used to get preferences for timezone and format
2502          * @param array $options Array, can contain the following keys:
2503          *   - 'timecorrection': time correction, can have the following values:
2504          *     - true: use user's preference
2505          *     - false: don't use time correction
2506          *     - int: value of time correction in minutes
2507          *   - 'format': format to use, can have the following values:
2508          *     - true: use user's preference
2509          *     - false: use default preference
2510          *     - string: format to use
2511          * @since 1.19
2512          * @return string
2513          */
2514         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2515                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2516         }
2517
2518         /**
2519          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2520          *
2521          * Determine the difference between the timestamp and the current time, and
2522          * generate a readable timestamp by returning "<N> <units> ago", where the
2523          * largest possible unit is used.
2524          *
2525          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2526          *
2527          * @param MWTimestamp $time
2528          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2529          * @param User|null $user User the timestamp is being generated for
2530          *  (or null to use main context's user)
2531          * @return string Formatted timestamp
2532          */
2533         public function getHumanTimestamp(
2534                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2535         ) {
2536                 if ( $relativeTo === null ) {
2537                         $relativeTo = new MWTimestamp();
2538                 }
2539                 if ( $user === null ) {
2540                         $user = RequestContext::getMain()->getUser();
2541                 }
2542
2543                 // Adjust for the user's timezone.
2544                 $offsetThis = $time->offsetForUser( $user );
2545                 $offsetRel = $relativeTo->offsetForUser( $user );
2546
2547                 $ts = '';
2548                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2549                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2550                 }
2551
2552                 // Reset the timezone on the objects.
2553                 $time->timestamp->sub( $offsetThis );
2554                 $relativeTo->timestamp->sub( $offsetRel );
2555
2556                 return $ts;
2557         }
2558
2559         /**
2560          * Convert an MWTimestamp into a pretty human-readable timestamp using
2561          * the given user preferences and relative base time.
2562          *
2563          * @see Language::getHumanTimestamp
2564          * @param MWTimestamp $ts Timestamp to prettify
2565          * @param MWTimestamp $relativeTo Base timestamp
2566          * @param User $user User preferences to use
2567          * @return string Human timestamp
2568          * @since 1.26
2569          */
2570         private function getHumanTimestampInternal(
2571                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2572         ) {
2573                 $diff = $ts->diff( $relativeTo );
2574                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2575                         (int)$relativeTo->timestamp->format( 'w' ) );
2576                 $days = $diff->days ?: (int)$diffDay;
2577                 if ( $diff->invert || $days > 5
2578                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2579                 ) {
2580                         // Timestamps are in different years: use full timestamp
2581                         // Also do full timestamp for future dates
2582                         /**
2583                          * @todo FIXME: Add better handling of future timestamps.
2584                          */
2585                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2586                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2587                 } elseif ( $days > 5 ) {
2588                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2589                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2590                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2591                 } elseif ( $days > 1 ) {
2592                         // Timestamp within the past week: show the day of the week and time
2593                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2594                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2595                         // Messages:
2596                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2597                         $ts = wfMessage( "$weekday-at" )
2598                                 ->inLanguage( $this )
2599                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2600                                 ->text();
2601                 } elseif ( $days == 1 ) {
2602                         // Timestamp was yesterday: say 'yesterday' and the time.
2603                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2604                         $ts = wfMessage( 'yesterday-at' )
2605                                 ->inLanguage( $this )
2606                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2607                                 ->text();
2608                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2609                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2610                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2611                         $ts = wfMessage( 'today-at' )
2612                                 ->inLanguage( $this )
2613                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2614                                 ->text();
2615
2616                 // From here on in, the timestamp was soon enough ago so that we can simply say
2617                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2618                 } elseif ( $diff->h == 1 ) {
2619                         // Less than 90 minutes, but more than an hour ago.
2620                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2621                 } elseif ( $diff->i >= 1 ) {
2622                         // A few minutes ago.
2623                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2624                 } elseif ( $diff->s >= 30 ) {
2625                         // Less than a minute, but more than 30 sec ago.
2626                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2627                 } else {
2628                         // Less than 30 seconds ago.
2629                         $ts = wfMessage( 'just-now' )->text();
2630                 }
2631
2632                 return $ts;
2633         }
2634
2635         /**
2636          * @param string $key
2637          * @return string|null
2638          */
2639         public function getMessage( $key ) {
2640                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2641         }
2642
2643         /**
2644          * @return array
2645          */
2646         function getAllMessages() {
2647                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2648         }
2649
2650         /**
2651          * @param string $in
2652          * @param string $out
2653          * @param string $string
2654          * @return string
2655          */
2656         public function iconv( $in, $out, $string ) {
2657                 # Even with //IGNORE iconv can whine about illegal characters in
2658                 # *input* string. We just ignore those too.
2659                 # REF: https://bugs.php.net/bug.php?id=37166
2660                 # REF: https://phabricator.wikimedia.org/T18885
2661                 Wikimedia\suppressWarnings();
2662                 $text = iconv( $in, $out . '//IGNORE', $string );
2663                 Wikimedia\restoreWarnings();
2664                 return $text;
2665         }
2666
2667         // callback functions for ucwords(), ucwordbreaks()
2668
2669         /**
2670          * @param array $matches
2671          * @return mixed|string
2672          */
2673         function ucwordbreaksCallbackAscii( $matches ) {
2674                 return $this->ucfirst( $matches[1] );
2675         }
2676
2677         /**
2678          * @param array $matches
2679          * @return string
2680          */
2681         function ucwordbreaksCallbackMB( $matches ) {
2682                 return mb_strtoupper( $matches[0] );
2683         }
2684
2685         /**
2686          * @param array $matches
2687          * @return string
2688          */
2689         function ucwordsCallbackMB( $matches ) {
2690                 return mb_strtoupper( $matches[0] );
2691         }
2692
2693         /**
2694          * Make a string's first character uppercase
2695          *
2696          * @param string $str
2697          *
2698          * @return string
2699          */
2700         public function ucfirst( $str ) {
2701                 $o = ord( $str );
2702                 if ( $o < 96 ) { // if already uppercase...
2703                         return $str;
2704                 } elseif ( $o < 128 ) {
2705                         return ucfirst( $str ); // use PHP's ucfirst()
2706                 } else {
2707                         // fall back to more complex logic in case of multibyte strings
2708                         return $this->uc( $str, true );
2709                 }
2710         }
2711
2712         /**
2713          * Convert a string to uppercase
2714          *
2715          * @param string $str
2716          * @param bool $first
2717          *
2718          * @return string
2719          */
2720         public function uc( $str, $first = false ) {
2721                 if ( $first ) {
2722                         if ( $this->isMultibyte( $str ) ) {
2723                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2724                         } else {
2725                                 return ucfirst( $str );
2726                         }
2727                 } else {
2728                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2729                 }
2730         }
2731
2732         /**
2733          * @param string $str
2734          * @return mixed|string
2735          */
2736         function lcfirst( $str ) {
2737                 $o = ord( $str );
2738                 if ( !$o ) {
2739                         return strval( $str );
2740                 } elseif ( $o >= 128 ) {
2741                         return $this->lc( $str, true );
2742                 } elseif ( $o > 96 ) {
2743                         return $str;
2744                 } else {
2745                         $str[0] = strtolower( $str[0] );
2746                         return $str;
2747                 }
2748         }
2749
2750         /**
2751          * @param string $str
2752          * @param bool $first
2753          * @return mixed|string
2754          */
2755         function lc( $str, $first = false ) {
2756                 if ( $first ) {
2757                         if ( $this->isMultibyte( $str ) ) {
2758                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2759                         } else {
2760                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2761                         }
2762                 } else {
2763                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2764                 }
2765         }
2766
2767         /**
2768          * @param string $str
2769          * @return bool
2770          */
2771         function isMultibyte( $str ) {
2772                 return strlen( $str ) !== mb_strlen( $str );
2773         }
2774
2775         /**
2776          * @param string $str
2777          * @return mixed|string
2778          */
2779         function ucwords( $str ) {
2780                 if ( $this->isMultibyte( $str ) ) {
2781                         $str = $this->lc( $str );
2782
2783                         // regexp to find first letter in each word (i.e. after each space)
2784                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2785
2786                         // function to use to capitalize a single char
2787                         return preg_replace_callback(
2788                                 $replaceRegexp,
2789                                 [ $this, 'ucwordsCallbackMB' ],
2790                                 $str
2791                         );
2792                 } else {
2793                         return ucwords( strtolower( $str ) );
2794                 }
2795         }
2796
2797         /**
2798          * capitalize words at word breaks
2799          *
2800          * @param string $str
2801          * @return mixed
2802          */
2803         function ucwordbreaks( $str ) {
2804                 if ( $this->isMultibyte( $str ) ) {
2805                         $str = $this->lc( $str );
2806
2807                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2808                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2809
2810                         // find first letter after word break
2811                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2812                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2813
2814                         return preg_replace_callback(
2815                                 $replaceRegexp,
2816                                 [ $this, 'ucwordbreaksCallbackMB' ],
2817                                 $str
2818                         );
2819                 } else {
2820                         return preg_replace_callback(
2821                                 '/\b([\w\x80-\xff]+)\b/',
2822                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2823                                 $str
2824                         );
2825                 }
2826         }
2827
2828         /**
2829          * Return a case-folded representation of $s
2830          *
2831          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2832          * and $s2 are the same except for the case of their characters. It is not
2833          * necessary for the value returned to make sense when displayed.
2834          *
2835          * Do *not* perform any other normalisation in this function. If a caller
2836          * uses this function when it should be using a more general normalisation
2837          * function, then fix the caller.
2838          *
2839          * @param string $s
2840          *
2841          * @return string
2842          */
2843         function caseFold( $s ) {
2844                 return $this->uc( $s );
2845         }
2846
2847         /**
2848          * @param string $s
2849          * @return string
2850          * @throws MWException
2851          */
2852         function checkTitleEncoding( $s ) {
2853                 if ( is_array( $s ) ) {
2854                         throw new MWException( 'Given array to checkTitleEncoding.' );
2855                 }
2856                 if ( StringUtils::isUtf8( $s ) ) {
2857                         return $s;
2858                 }
2859
2860                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2861         }
2862
2863         /**
2864          * @return array
2865          */
2866         function fallback8bitEncoding() {
2867                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2868         }
2869
2870         /**
2871          * Most writing systems use whitespace to break up words.
2872          * Some languages such as Chinese don't conventionally do this,
2873          * which requires special handling when breaking up words for
2874          * searching etc.
2875          *
2876          * @return bool
2877          */
2878         function hasWordBreaks() {
2879                 return true;
2880         }
2881
2882         /**
2883          * Some languages such as Chinese require word segmentation,
2884          * Specify such segmentation when overridden in derived class.
2885          *
2886          * @param string $string
2887          * @return string
2888          */
2889         function segmentByWord( $string ) {
2890                 return $string;
2891         }
2892
2893         /**
2894          * Some languages have special punctuation need to be normalized.
2895          * Make such changes here.
2896          *
2897          * @param string $string
2898          * @return string
2899          */
2900         function normalizeForSearch( $string ) {
2901                 return self::convertDoubleWidth( $string );
2902         }
2903
2904         /**
2905          * convert double-width roman characters to single-width.
2906          * range: ff00-ff5f ~= 0020-007f
2907          *
2908          * @param string $string
2909          *
2910          * @return string
2911          */
2912         protected static function convertDoubleWidth( $string ) {
2913                 static $full = null;
2914                 static $half = null;
2915
2916                 if ( $full === null ) {
2917                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2918                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2919                         $full = str_split( $fullWidth, 3 );
2920                         $half = str_split( $halfWidth );
2921                 }
2922
2923                 $string = str_replace( $full, $half, $string );
2924                 return $string;
2925         }
2926
2927         /**
2928          * @param string $string
2929          * @param string $pattern
2930          * @return string
2931          */
2932         protected static function insertSpace( $string, $pattern ) {
2933                 $string = preg_replace( $pattern, " $1 ", $string );
2934                 $string = preg_replace( '/ +/', ' ', $string );
2935                 return $string;
2936         }
2937
2938         /**
2939          * @param array $termsArray
2940          * @return array
2941          */
2942         function convertForSearchResult( $termsArray ) {
2943                 # some languages, e.g. Chinese, need to do a conversion
2944                 # in order for search results to be displayed correctly
2945                 return $termsArray;
2946         }
2947
2948         /**
2949          * Get the first character of a string.
2950          *
2951          * @param string $s
2952          * @return string
2953          */
2954         function firstChar( $s ) {
2955                 $matches = [];
2956                 preg_match(
2957                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2958                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2959                         $s,
2960                         $matches
2961                 );
2962
2963                 if ( isset( $matches[1] ) ) {
2964                         if ( strlen( $matches[1] ) != 3 ) {
2965                                 return $matches[1];
2966                         }
2967
2968                         // Break down Hangul syllables to grab the first jamo
2969                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2970                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2971                                 return $matches[1];
2972                         } elseif ( $code < 0xb098 ) {
2973                                 return "\u{3131}";
2974                         } elseif ( $code < 0xb2e4 ) {
2975                                 return "\u{3134}";
2976                         } elseif ( $code < 0xb77c ) {
2977                                 return "\u{3137}";
2978                         } elseif ( $code < 0xb9c8 ) {
2979                                 return "\u{3139}";
2980                         } elseif ( $code < 0xbc14 ) {
2981                                 return "\u{3141}";
2982                         } elseif ( $code < 0xc0ac ) {
2983                                 return "\u{3142}";
2984                         } elseif ( $code < 0xc544 ) {
2985                                 return "\u{3145}";
2986                         } elseif ( $code < 0xc790 ) {
2987                                 return "\u{3147}";
2988                         } elseif ( $code < 0xcc28 ) {
2989                                 return "\u{3148}";
2990                         } elseif ( $code < 0xce74 ) {
2991                                 return "\u{314A}";
2992                         } elseif ( $code < 0xd0c0 ) {
2993                                 return "\u{314B}";
2994                         } elseif ( $code < 0xd30c ) {
2995                                 return "\u{314C}";
2996                         } elseif ( $code < 0xd558 ) {
2997                                 return "\u{314D}";
2998                         } else {
2999                                 return "\u{314E}";
3000                         }
3001                 } else {
3002                         return '';
3003                 }
3004         }
3005
3006         /**
3007          * @deprecated No-op since 1.28
3008          */
3009         function initEncoding() {
3010                 wfDeprecated( __METHOD__, '1.28' );
3011                 // No-op.
3012         }
3013
3014         /**
3015          * @param string $s
3016          * @return string
3017          * @deprecated No-op since 1.28
3018          */
3019         function recodeForEdit( $s ) {
3020                 wfDeprecated( __METHOD__, '1.28' );
3021                 return $s;
3022         }
3023
3024         /**
3025          * @param string $s
3026          * @return string
3027          * @deprecated No-op since 1.28
3028          */
3029         function recodeInput( $s ) {
3030                 wfDeprecated( __METHOD__, '1.28' );
3031                 return $s;
3032         }
3033
3034         /**
3035          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3036          * also cleans up certain backwards-compatible sequences, converting them
3037          * to the modern Unicode equivalent.
3038          *
3039          * This is language-specific for performance reasons only.
3040          *
3041          * @param string $s
3042          *
3043          * @return string
3044          */
3045         public function normalize( $s ) {
3046                 global $wgAllUnicodeFixes;
3047                 $s = UtfNormal\Validator::cleanUp( $s );
3048                 if ( $wgAllUnicodeFixes ) {
3049                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3050                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3051                 }
3052
3053                 return $s;
3054         }
3055
3056         /**
3057          * Transform a string using serialized data stored in the given file (which
3058          * must be in the serialized subdirectory of $IP). The file contains pairs
3059          * mapping source characters to destination characters.
3060          *
3061          * The data is cached in process memory. This will go faster if you have the
3062          * FastStringSearch extension.
3063          *
3064          * @param string $file
3065          * @param string $string
3066          *
3067          * @throws MWException
3068          * @return string
3069          */
3070         protected function transformUsingPairFile( $file, $string ) {
3071                 if ( !isset( $this->transformData[$file] ) ) {
3072                         global $IP;
3073                         $data = require "$IP/languages/data/{$file}";
3074                         $this->transformData[$file] = new ReplacementArray( $data );
3075                 }
3076                 return $this->transformData[$file]->replace( $string );
3077         }
3078
3079         /**
3080          * For right-to-left language support
3081          *
3082          * @return bool
3083          */
3084         function isRTL() {
3085                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3086         }
3087
3088         /**
3089          * Return the correct HTML 'dir' attribute value for this language.
3090          * @return string
3091          */
3092         function getDir() {
3093                 return $this->isRTL() ? 'rtl' : 'ltr';
3094         }
3095
3096         /**
3097          * Return 'left' or 'right' as appropriate alignment for line-start
3098          * for this language's text direction.
3099          *
3100          * Should be equivalent to CSS3 'start' text-align value....
3101          *
3102          * @return string
3103          */
3104         function alignStart() {
3105                 return $this->isRTL() ? 'right' : 'left';
3106         }
3107
3108         /**
3109          * Return 'right' or 'left' as appropriate alignment for line-end
3110          * for this language's text direction.
3111          *
3112          * Should be equivalent to CSS3 'end' text-align value....
3113          *
3114          * @return string
3115          */
3116         function alignEnd() {
3117                 return $this->isRTL() ? 'left' : 'right';
3118         }
3119
3120         /**
3121          * A hidden direction mark (LRM or RLM), depending on the language direction.
3122          * Unlike getDirMark(), this function returns the character as an HTML entity.
3123          * This function should be used when the output is guaranteed to be HTML,
3124          * because it makes the output HTML source code more readable. When
3125          * the output is plain text or can be escaped, getDirMark() should be used.
3126          *
3127          * @param bool $opposite Get the direction mark opposite to your language
3128          * @return string
3129          * @since 1.20
3130          */
3131         function getDirMarkEntity( $opposite = false ) {
3132                 if ( $opposite ) {
3133                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3134                 }
3135                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3136         }
3137
3138         /**
3139          * A hidden direction mark (LRM or RLM), depending on the language direction.
3140          * This function produces them as invisible Unicode characters and
3141          * the output may be hard to read and debug, so it should only be used
3142          * when the output is plain text or can be escaped. When the output is
3143          * HTML, use getDirMarkEntity() instead.
3144          *
3145          * @param bool $opposite Get the direction mark opposite to your language
3146          * @return string
3147          */
3148         function getDirMark( $opposite = false ) {
3149                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3150                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3151                 if ( $opposite ) {
3152                         return $this->isRTL() ? $lrm : $rlm;
3153                 }
3154                 return $this->isRTL() ? $rlm : $lrm;
3155         }
3156
3157         /**
3158          * @return array
3159          */
3160         function capitalizeAllNouns() {
3161                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3162         }
3163
3164         /**
3165          * An arrow, depending on the language direction.
3166          *
3167          * @param string $direction The direction of the arrow: forwards (default),
3168          *   backwards, left, right, up, down.
3169          * @return string
3170          */
3171         function getArrow( $direction = 'forwards' ) {
3172                 switch ( $direction ) {
3173                         case 'forwards':
3174                                 return $this->isRTL() ? '←' : '→';
3175                         case 'backwards':
3176                                 return $this->isRTL() ? '→' : '←';
3177                         case 'left':
3178                                 return '←';
3179                         case 'right':
3180                                 return '→';
3181                         case 'up':
3182                                 return '↑';
3183                         case 'down':
3184                                 return '↓';
3185                 }
3186         }
3187
3188         /**
3189          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3190          *
3191          * @return bool
3192          */
3193         function linkPrefixExtension() {
3194                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3195         }
3196
3197         /**
3198          * Get all magic words from cache.
3199          * @return array
3200          */
3201         function getMagicWords() {
3202                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3203         }
3204
3205         /**
3206          * Fill a MagicWord object with data from here
3207          *
3208          * @param MagicWord $mw
3209          */
3210         function getMagic( $mw ) {
3211                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3212                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3213
3214                 if ( !is_array( $rawEntry ) ) {
3215                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3216                 } else {
3217                         $mw->mCaseSensitive = $rawEntry[0];
3218                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3219                 }
3220         }
3221
3222         /**
3223          * Add magic words to the extension array
3224          *
3225          * @param array $newWords
3226          */
3227         function addMagicWordsByLang( $newWords ) {
3228                 $fallbackChain = $this->getFallbackLanguages();
3229                 $fallbackChain = array_reverse( $fallbackChain );
3230                 foreach ( $fallbackChain as $code ) {
3231                         if ( isset( $newWords[$code] ) ) {
3232                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3233                         }
3234                 }
3235         }
3236
3237         /**
3238          * Get special page names, as an associative array
3239          *   canonical name => array of valid names, including aliases
3240          * @return array
3241          */
3242         function getSpecialPageAliases() {
3243                 // Cache aliases because it may be slow to load them
3244                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3245                         // Initialise array
3246                         $this->mExtendedSpecialPageAliases =
3247                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3248                 }
3249
3250                 return $this->mExtendedSpecialPageAliases;
3251         }
3252
3253         /**
3254          * Italic is unsuitable for some languages
3255          *
3256          * @param string $text The text to be emphasized.
3257          * @return string
3258          */
3259         function emphasize( $text ) {
3260                 return "<em>$text</em>";
3261         }
3262
3263         /**
3264          * Normally we output all numbers in plain en_US style, that is
3265          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3266          * point twohundredthirtyfive. However this is not suitable for all
3267          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3268          * Icelandic just want to use commas instead of dots, and dots instead
3269          * of commas like "293.291,235".
3270          *
3271          * An example of this function being called:
3272          * <code>
3273          * wfMessage( 'message' )->numParams( $num )->text()
3274          * </code>
3275          *
3276          * See $separatorTransformTable on MessageIs.php for
3277          * the , => . and . => , implementation.
3278          *
3279          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3280          * @param int|float $number The string to be formatted, should be an integer
3281          *   or a floating point number.
3282          * @param bool $nocommafy Set to true for special numbers like dates
3283          * @return string
3284          */
3285         public function formatNum( $number, $nocommafy = false ) {
3286                 global $wgTranslateNumerals;
3287                 if ( !$nocommafy ) {
3288                         $number = $this->commafy( $number );
3289                         $s = $this->separatorTransformTable();
3290                         if ( $s ) {
3291                                 $number = strtr( $number, $s );
3292                         }
3293                 }
3294
3295                 if ( $wgTranslateNumerals ) {
3296                         $s = $this->digitTransformTable();
3297                         if ( $s ) {
3298                                 $number = strtr( $number, $s );
3299                         }
3300                 }
3301
3302                 return (string)$number;
3303         }
3304
3305         /**
3306          * Front-end for non-commafied formatNum
3307          *
3308          * @param int|float $number The string to be formatted, should be an integer
3309          *        or a floating point number.
3310          * @since 1.21
3311          * @return string
3312          */
3313         public function formatNumNoSeparators( $number ) {
3314                 return $this->formatNum( $number, true );
3315         }
3316
3317         /**
3318          * @param string $number
3319          * @return string
3320          */
3321         public function parseFormattedNumber( $number ) {
3322                 $s = $this->digitTransformTable();
3323                 if ( $s ) {
3324                         // eliminate empty array values such as ''. (T66347)
3325                         $s = array_filter( $s );
3326                         $number = strtr( $number, array_flip( $s ) );
3327                 }
3328
3329                 $s = $this->separatorTransformTable();
3330                 if ( $s ) {
3331                         // eliminate empty array values such as ''. (T66347)
3332                         $s = array_filter( $s );
3333                         $number = strtr( $number, array_flip( $s ) );
3334                 }
3335
3336                 $number = strtr( $number, [ ',' => '' ] );
3337                 return $number;
3338         }
3339
3340         /**
3341          * Adds commas to a given number
3342          * @since 1.19
3343          * @param mixed $number
3344          * @return string
3345          */
3346         function commafy( $number ) {
3347                 $digitGroupingPattern = $this->digitGroupingPattern();
3348                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3349                 if ( $number === null ) {
3350                         return '';
3351                 }
3352
3353                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3354                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3355                         // In some languages it's conventional not to insert a thousands separator
3356                         // in numbers that are four digits long (1000-9999).
3357                         if ( $minimumGroupingDigits ) {
3358                                 // Number of '#' characters after last comma in the grouping pattern.
3359                                 // The pattern is hardcoded here, but this would vary for different patterns.
3360                                 $primaryGroupingSize = 3;
3361                                 // Maximum length of a number to suppress digit grouping for.
3362                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3363                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3364                                         return $number;
3365                                 }
3366                         }
3367                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3368                 } else {
3369                         // Ref: http://cldr.unicode.org/translation/number-patterns
3370                         $sign = "";
3371                         if ( intval( $number ) < 0 ) {
3372                                 // For negative numbers apply the algorithm like positive number and add sign.
3373                                 $sign = "-";
3374                                 $number = substr( $number, 1 );
3375                         }
3376                         $integerPart = [];
3377                         $decimalPart = [];
3378                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3379                         preg_match( "/\d+/", $number, $integerPart );
3380                         preg_match( "/\.\d*/", $number, $decimalPart );
3381                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3382                         if ( $groupedNumber === $number ) {
3383                                 // the string does not have any number part. Eg: .12345
3384                                 return $sign . $groupedNumber;
3385                         }
3386                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3387                         while ( $start > 0 ) {
3388                                 $match = $matches[0][$numMatches - 1];
3389                                 $matchLen = strlen( $match );
3390                                 $start = $end - $matchLen;
3391                                 if ( $start < 0 ) {
3392                                         $start = 0;
3393                                 }
3394                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3395                                 $end = $start;
3396                                 if ( $numMatches > 1 ) {
3397                                         // use the last pattern for the rest of the number
3398                                         $numMatches--;
3399                                 }
3400                                 if ( $start > 0 ) {
3401                                         $groupedNumber = "," . $groupedNumber;
3402                                 }
3403                         }
3404                         return $sign . $groupedNumber;
3405                 }
3406         }
3407
3408         /**
3409          * @return string
3410          */
3411         function digitGroupingPattern() {
3412                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3413         }
3414
3415         /**
3416          * @return array
3417          */
3418         function digitTransformTable() {
3419                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3420         }
3421
3422         /**
3423          * @return array
3424          */
3425         function separatorTransformTable() {
3426                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3427         }
3428
3429         /**
3430          * @return int|null
3431          */
3432         function minimumGroupingDigits() {
3433                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3434         }
3435
3436         /**
3437          * Take a list of strings and build a locale-friendly comma-separated
3438          * list, using the local comma-separator message.
3439          * The last two strings are chained with an "and".
3440          *
3441          * @param string[] $list
3442          * @return string
3443          */
3444         public function listToText( array $list ) {
3445                 $itemCount = count( $list );
3446                 if ( $itemCount < 1 ) {
3447                         return '';
3448                 }
3449                 $text = array_pop( $list );
3450                 if ( $itemCount > 1 ) {
3451                         $and = $this->msg( 'and' )->escaped();
3452                         $space = $this->msg( 'word-separator' )->escaped();
3453                         $comma = '';
3454                         if ( $itemCount > 2 ) {
3455                                 $comma = $this->msg( 'comma-separator' )->escaped();
3456                         }
3457                         $text = implode( $comma, $list ) . $and . $space . $text;
3458                 }
3459                 return $text;
3460         }
3461
3462         /**
3463          * Take a list of strings and build a locale-friendly comma-separated
3464          * list, using the local comma-separator message.
3465          * @param string[] $list Array of strings to put in a comma list
3466          * @return string
3467          */
3468         function commaList( array $list ) {
3469                 return implode(
3470                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3471                         $list
3472                 );
3473         }
3474
3475         /**
3476          * Take a list of strings and build a locale-friendly semicolon-separated
3477          * list, using the local semicolon-separator message.
3478          * @param string[] $list Array of strings to put in a semicolon list
3479          * @return string
3480          */
3481         function semicolonList( array $list ) {
3482                 return implode(
3483                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3484                         $list
3485                 );
3486         }
3487
3488         /**
3489          * Same as commaList, but separate it with the pipe instead.
3490          * @param string[] $list Array of strings to put in a pipe list
3491          * @return string
3492          */
3493         function pipeList( array $list ) {
3494                 return implode(
3495                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3496                         $list
3497                 );
3498         }
3499
3500         /**
3501          * Truncate a string to a specified length in bytes, appending an optional
3502          * string (e.g. for ellipsis)
3503          *
3504          * If $length is negative, the string will be truncated from the beginning
3505          *
3506          * @since 1.31
3507          *
3508          * @param string $string String to truncate
3509          * @param int $length Maximum length in bytes
3510          * @param string $ellipsis String to append to the end of truncated text
3511          * @param bool $adjustLength Subtract length of ellipsis from $length
3512          *
3513          * @return string
3514          */
3515         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3516                 return $this->truncateInternal(
3517                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3518                 );
3519         }
3520
3521         /**
3522          * Truncate a string to a specified number of characters, appending an optional
3523          * string (e.g. for ellipsis).
3524          *
3525          * This provides multibyte version of truncate() method of this class, suitable for truncation
3526          * based on number of characters, instead of number of bytes.
3527          *
3528          * If $length is negative, the string will be truncated from the beginning.
3529          *
3530          * @since 1.31
3531          *
3532          * @param string $string String to truncate
3533          * @param int $length Maximum number of characters
3534          * @param string $ellipsis String to append to the end of truncated text
3535          * @param bool $adjustLength Subtract length of ellipsis from $length
3536          *
3537          * @return string
3538          */
3539         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3540                 // Passing encoding to mb_strlen and mb_substr is optional.
3541                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3542                 // explicit specification of encoding is skipped.
3543                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3544                 return $this->truncateInternal(
3545                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3546                 );
3547         }
3548
3549         /**
3550          * Internal method used for truncation. This method abstracts text truncation into
3551          * one common method, allowing users to provide length measurement function and
3552          * function for finding substring.
3553          *
3554          * For usages, see truncateForDatabase and truncateForVisual.
3555          *
3556          * @param string $string String to truncate
3557          * @param int $length Maximum length of final text
3558          * @param string $ellipsis String to append to the end of truncated text
3559          * @param bool $adjustLength Subtract length of ellipsis from $length
3560          * @param callable $measureLength Callable function used for determining the length of text
3561          * @param callable $getSubstring Callable function used for getting the substrings
3562          *
3563          * @return string
3564          */
3565         private function truncateInternal(
3566                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3567         ) {
3568                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3569                         throw new InvalidArgumentException( 'Invalid callback provided' );
3570                 }
3571
3572                 # Check if there is no need to truncate
3573                 if ( $measureLength( $string ) <= abs( $length ) ) {
3574                         return $string; // no need to truncate
3575                 }
3576
3577                 # Use the localized ellipsis character
3578                 if ( $ellipsis == '...' ) {
3579                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3580                 }
3581                 if ( $length == 0 ) {
3582                         return $ellipsis; // convention
3583                 }
3584
3585                 $stringOriginal = $string;
3586                 # If ellipsis length is >= $length then we can't apply $adjustLength
3587                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3588                         $string = $ellipsis; // this can be slightly unexpected
3589                 # Otherwise, truncate and add ellipsis...
3590                 } else {
3591                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3592                         if ( $length > 0 ) {
3593                                 $length -= $ellipsisLength;
3594                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3595                                 $string = $this->removeBadCharLast( $string );
3596                                 $string = rtrim( $string );
3597                                 $string = $string . $ellipsis;
3598                         } else {
3599                                 $length += $ellipsisLength;
3600                                 $string = $getSubstring( $string, $length ); // ...xyz
3601                                 $string = $this->removeBadCharFirst( $string );
3602                                 $string = ltrim( $string );
3603                                 $string = $ellipsis . $string;
3604                         }
3605                 }
3606
3607                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3608                 # This check is *not* redundant if $adjustLength, due to the single case where
3609                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3610                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3611                         return $string;
3612                 } else {
3613                         return $stringOriginal;
3614                 }
3615         }
3616
3617         /**
3618          * Remove bytes that represent an incomplete Unicode character
3619          * at the end of string (e.g. bytes of the char are missing)
3620          *
3621          * @param string $string
3622          * @return string
3623          */
3624         protected function removeBadCharLast( $string ) {
3625                 if ( $string != '' ) {
3626                         $char = ord( $string[strlen( $string ) - 1] );
3627                         $m = [];
3628                         if ( $char >= 0xc0 ) {
3629                                 # We got the first byte only of a multibyte char; remove it.
3630                                 $string = substr( $string, 0, -1 );
3631                         } elseif ( $char >= 0x80 &&
3632                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3633                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3634                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3635                         ) {
3636                                 # We chopped in the middle of a character; remove it
3637                                 $string = $m[1];
3638                         }
3639                 }
3640                 return $string;
3641         }
3642
3643         /**
3644          * Remove bytes that represent an incomplete Unicode character
3645          * at the start of string (e.g. bytes of the char are missing)
3646          *
3647          * @param string $string
3648          * @return string
3649          */
3650         protected function removeBadCharFirst( $string ) {
3651                 if ( $string != '' ) {
3652                         $char = ord( $string[0] );
3653                         if ( $char >= 0x80 && $char < 0xc0 ) {
3654                                 # We chopped in the middle of a character; remove the whole thing
3655                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3656                         }
3657                 }
3658                 return $string;
3659         }
3660
3661         /**
3662          * Truncate a string of valid HTML to a specified length in bytes,
3663          * appending an optional string (e.g. for ellipses), and return valid HTML
3664          *
3665          * This is only intended for styled/linked text, such as HTML with
3666          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3667          * Also, this will not detect things like "display:none" CSS.
3668          *
3669          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3670          *
3671          * @param string $text HTML string to truncate
3672          * @param int $length (zero/positive) Maximum length (including ellipses)
3673          * @param string $ellipsis String to append to the truncated text
3674          * @return string
3675          */
3676         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3677                 # Use the localized ellipsis character
3678                 if ( $ellipsis == '...' ) {
3679                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3680                 }
3681                 # Check if there is clearly no need to truncate
3682                 if ( $length <= 0 ) {
3683                         return $ellipsis; // no text shown, nothing to format (convention)
3684                 } elseif ( strlen( $text ) <= $length ) {
3685                         return $text; // string short enough even *with* HTML (short-circuit)
3686                 }
3687
3688                 $dispLen = 0; // innerHTML legth so far
3689                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3690                 $tagType = 0; // 0-open, 1-close
3691                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3692                 $entityState = 0; // 0-not entity, 1-entity
3693                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3694                 $openTags = []; // open tag stack
3695                 $maybeState = null; // possible truncation state
3696
3697                 $textLen = strlen( $text );
3698                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3699                 for ( $pos = 0; true; ++$pos ) {
3700                         # Consider truncation once the display length has reached the maximim.
3701                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3702                         # Check that we're not in the middle of a bracket/entity...
3703                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3704                                 if ( !$testingEllipsis ) {
3705                                         $testingEllipsis = true;
3706                                         # Save where we are; we will truncate here unless there turn out to
3707                                         # be so few remaining characters that truncation is not necessary.
3708                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3709                                                 $maybeState = [ $ret, $openTags ]; // save state
3710                                         }
3711                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3712                                         # String in fact does need truncation, the truncation point was OK.
3713                                         list( $ret, $openTags ) = $maybeState; // reload state
3714                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3715                                         $ret .= $ellipsis; // add ellipsis
3716                                         break;
3717                                 }
3718                         }
3719                         if ( $pos >= $textLen ) {
3720                                 break; // extra iteration just for above checks
3721                         }
3722
3723                         # Read the next char...
3724                         $ch = $text[$pos];
3725                         $lastCh = $pos ? $text[$pos - 1] : '';
3726                         $ret .= $ch; // add to result string
3727                         if ( $ch == '<' ) {
3728                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3729                                 $entityState = 0; // for bad HTML
3730                                 $bracketState = 1; // tag started (checking for backslash)
3731                         } elseif ( $ch == '>' ) {
3732                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3733                                 $entityState = 0; // for bad HTML
3734                                 $bracketState = 0; // out of brackets
3735                         } elseif ( $bracketState == 1 ) {
3736                                 if ( $ch == '/' ) {
3737                                         $tagType = 1; // close tag (e.g. "</span>")
3738                                 } else {
3739                                         $tagType = 0; // open tag (e.g. "<span>")
3740                                         $tag .= $ch;
3741                                 }
3742                                 $bracketState = 2; // building tag name
3743                         } elseif ( $bracketState == 2 ) {
3744                                 if ( $ch != ' ' ) {
3745                                         $tag .= $ch;
3746                                 } else {
3747                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3748                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3749                                 }
3750                         } elseif ( $bracketState == 0 ) {
3751                                 if ( $entityState ) {
3752                                         if ( $ch == ';' ) {
3753                                                 $entityState = 0;
3754                                                 $dispLen++; // entity is one displayed char
3755                                         }
3756                                 } else {
3757                                         if ( $neLength == 0 && !$maybeState ) {
3758                                                 // Save state without $ch. We want to *hit* the first
3759                                                 // display char (to get tags) but not *use* it if truncating.
3760                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3761                                         }
3762                                         if ( $ch == '&' ) {
3763                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3764                                         } else {
3765                                                 $dispLen++; // this char is displayed
3766                                                 // Add the next $max display text chars after this in one swoop...
3767                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3768                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3769                                                 $dispLen += $skipped;
3770                                                 $pos += $skipped;
3771                                         }
3772                                 }
3773                         }
3774                 }
3775                 // Close the last tag if left unclosed by bad HTML
3776                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3777                 while ( count( $openTags ) > 0 ) {
3778                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3779                 }
3780                 return $ret;
3781         }
3782
3783         /**
3784          * truncateHtml() helper function
3785          * like strcspn() but adds the skipped chars to $ret
3786          *
3787          * @param string $ret
3788          * @param string $text
3789          * @param string $search
3790          * @param int $start
3791          * @param null|int $len
3792          * @return int
3793          */
3794         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3795                 if ( $len === null ) {
3796                         $len = -1; // -1 means "no limit" for strcspn
3797                 } elseif ( $len < 0 ) {
3798                         $len = 0; // sanity
3799                 }
3800                 $skipCount = 0;
3801                 if ( $start < strlen( $text ) ) {
3802                         $skipCount = strcspn( $text, $search, $start, $len );
3803                         $ret .= substr( $text, $start, $skipCount );
3804                 }
3805                 return $skipCount;
3806         }
3807
3808         /**
3809          * truncateHtml() helper function
3810          * (a) push or pop $tag from $openTags as needed
3811          * (b) clear $tag value
3812          * @param string &$tag Current HTML tag name we are looking at
3813          * @param int $tagType (0-open tag, 1-close tag)
3814          * @param string $lastCh Character before the '>' that ended this tag
3815          * @param array &$openTags Open tag stack (not accounting for $tag)
3816          */
3817         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3818                 $tag = ltrim( $tag );
3819                 if ( $tag != '' ) {
3820                         if ( $tagType == 0 && $lastCh != '/' ) {
3821                                 $openTags[] = $tag; // tag opened (didn't close itself)
3822                         } elseif ( $tagType == 1 ) {
3823                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3824                                         array_pop( $openTags ); // tag closed
3825                                 }
3826                         }
3827                         $tag = '';
3828                 }
3829         }
3830
3831         /**
3832          * Grammatical transformations, needed for inflected languages
3833          * Invoked by putting {{grammar:case|word}} in a message
3834          *
3835          * @param string $word
3836          * @param string $case
3837          * @return string
3838          */
3839         function convertGrammar( $word, $case ) {
3840                 global $wgGrammarForms;
3841                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3842                         return $wgGrammarForms[$this->getCode()][$case][$word];
3843                 }
3844
3845                 $grammarTransformations = $this->getGrammarTransformations();
3846
3847                 if ( isset( $grammarTransformations[$case] ) ) {
3848                         $forms = $grammarTransformations[$case];
3849
3850                         // Some names of grammar rules are aliases for other rules.
3851                         // In such cases the value is a string rather than object,
3852                         // so load the actual rules.
3853                         if ( is_string( $forms ) ) {
3854                                 $forms = $grammarTransformations[$forms];
3855                         }
3856
3857                         foreach ( array_values( $forms ) as $rule ) {
3858                                 $form = $rule[0];
3859
3860                                 if ( $form === '@metadata' ) {
3861                                         continue;
3862                                 }
3863
3864                                 $replacement = $rule[1];
3865
3866                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3867                                 $patternMatches = preg_match( $regex, $word );
3868
3869                                 if ( $patternMatches === false ) {
3870                                         wfLogWarning(
3871                                                 'An error occurred while processing grammar. ' .
3872                                                 "Word: '$word'. Regex: /$form/."
3873                                         );
3874                                 } elseif ( $patternMatches === 1 ) {
3875                                         $word = preg_replace( $regex, $replacement, $word );
3876
3877                                         break;
3878                                 }
3879                         }
3880                 }
3881
3882                 return $word;
3883         }
3884
3885         /**
3886          * Get the grammar forms for the content language
3887          * @return array Array of grammar forms
3888          * @since 1.20
3889          */
3890         function getGrammarForms() {
3891                 global $wgGrammarForms;
3892                 if ( isset( $wgGrammarForms[$this->getCode()] )
3893                         && is_array( $wgGrammarForms[$this->getCode()] )
3894                 ) {
3895                         return $wgGrammarForms[$this->getCode()];
3896                 }
3897
3898                 return [];
3899         }
3900
3901         /**
3902          * Get the grammar transformations data for the language.
3903          * Used like grammar forms, with {{GRAMMAR}} and cases,
3904          * but uses pairs of regexes and replacements instead of code.
3905          *
3906          * @return array[] Array of grammar transformations.
3907          * @throws MWException
3908          * @since 1.28
3909          */
3910         public function getGrammarTransformations() {
3911                 $languageCode = $this->getCode();
3912
3913                 if ( self::$grammarTransformations === null ) {
3914                         self::$grammarTransformations = new MapCacheLRU( 10 );
3915                 }
3916
3917                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3918                         return self::$grammarTransformations->get( $languageCode );
3919                 }
3920
3921                 $data = [];
3922
3923                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3924                 if ( is_readable( $grammarDataFile ) ) {
3925                         $data = FormatJson::decode(
3926                                 file_get_contents( $grammarDataFile ),
3927                                 true
3928                         );
3929
3930                         if ( $data === null ) {
3931                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3932                         }
3933
3934                         self::$grammarTransformations->set( $languageCode, $data );
3935                 }
3936
3937                 return $data;
3938         }
3939
3940         /**
3941          * Provides an alternative text depending on specified gender.
3942          * Usage {{gender:username|masculine|feminine|unknown}}.
3943          * username is optional, in which case the gender of current user is used,
3944          * but only in (some) interface messages; otherwise default gender is used.
3945          *
3946          * If no forms are given, an empty string is returned. If only one form is
3947          * given, it will be returned unconditionally. These details are implied by
3948          * the caller and cannot be overridden in subclasses.
3949          *
3950          * If three forms are given, the default is to use the third (unknown) form.
3951          * If fewer than three forms are given, the default is to use the first (masculine) form.
3952          * These details can be overridden in subclasses.
3953          *
3954          * @param string $gender
3955          * @param array $forms
3956          *
3957          * @return string
3958          */
3959         function gender( $gender, $forms ) {
3960                 if ( !count( $forms ) ) {
3961                         return '';
3962                 }
3963                 $forms = $this->preConvertPlural( $forms, 2 );
3964                 if ( $gender === 'male' ) {
3965                         return $forms[0];
3966                 }
3967                 if ( $gender === 'female' ) {
3968                         return $forms[1];
3969                 }
3970                 return $forms[2] ?? $forms[0];
3971         }
3972
3973         /**
3974          * Plural form transformations, needed for some languages.
3975          * For example, there are 3 form of plural in Russian and Polish,
3976          * depending on "count mod 10". See [[w:Plural]]
3977          * For English it is pretty simple.
3978          *
3979          * Invoked by putting {{plural:count|wordform1|wordform2}}
3980          * or {{plural:count|wordform1|wordform2|wordform3}}
3981          *
3982          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3983          *
3984          * @param int $count Non-localized number
3985          * @param array $forms Different plural forms
3986          * @return string Correct form of plural for $count in this language
3987          */
3988         function convertPlural( $count, $forms ) {
3989                 // Handle explicit n=pluralform cases
3990                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3991                 if ( is_string( $forms ) ) {
3992                         return $forms;
3993                 }
3994                 if ( !count( $forms ) ) {
3995                         return '';
3996                 }
3997
3998                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3999                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4000                 return $forms[$pluralForm];
4001         }
4002
4003         /**
4004          * Handles explicit plural forms for Language::convertPlural()
4005          *
4006          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4007          * If an explicitly defined plural form matches the $count, then
4008          * string value returned, otherwise array returned for further consideration
4009          * by CLDR rules or overridden convertPlural().
4010          *
4011          * @since 1.23
4012          *
4013          * @param int $count Non-localized number
4014          * @param array $forms Different plural forms
4015          *
4016          * @return array|string
4017          */
4018         protected function handleExplicitPluralForms( $count, array $forms ) {
4019                 foreach ( $forms as $index => $form ) {
4020                         if ( preg_match( '/\d+=/i', $form ) ) {
4021                                 $pos = strpos( $form, '=' );
4022                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4023                                         return substr( $form, $pos + 1 );
4024                                 }
4025                                 unset( $forms[$index] );
4026                         }
4027                 }
4028                 return array_values( $forms );
4029         }
4030
4031         /**
4032          * Checks that convertPlural was given an array and pads it to requested
4033          * amount of forms by copying the last one.
4034          *
4035          * @param array $forms Array of forms given to convertPlural
4036          * @param int $count How many forms should there be at least
4037          * @return array Padded array of forms or an exception if not an array
4038          */
4039         protected function preConvertPlural( /* Array */ $forms, $count ) {
4040                 while ( count( $forms ) < $count ) {
4041                         $forms[] = $forms[count( $forms ) - 1];
4042                 }
4043                 return $forms;
4044         }
4045
4046         /**
4047          * Wraps argument with unicode control characters for directionality safety
4048          *
4049          * This solves the problem where directionality-neutral characters at the edge of
4050          * the argument string get interpreted with the wrong directionality from the
4051          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4052          *
4053          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4054          * directionality of the argument string, using the BIDI algorithm's own "First
4055          * strong directional codepoint" rule. Essentially, this works round the fact that
4056          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4057          * direction inference). The latter is cleaner but still not widely supported.
4058          *
4059          * @param string $text Text to wrap
4060          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4061          */
4062         public function embedBidi( $text = '' ) {
4063                 $dir = self::strongDirFromContent( $text );
4064                 if ( $dir === 'ltr' ) {
4065                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4066                         return self::$lre . $text . self::$pdf;
4067                 }
4068                 if ( $dir === 'rtl' ) {
4069                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4070                         return self::$rle . $text . self::$pdf;
4071                 }
4072                 // No strong directionality: do not wrap
4073                 return $text;
4074         }
4075
4076         /**
4077          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4078          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4079          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4080          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4081          * match up with it.
4082          *
4083          * @param string $str The validated block duration in English
4084          * @param User|null $user User object to use timezone from or null for $wgUser
4085          * @param int $now Current timestamp, for formatting relative block durations
4086          * @return string Somehow translated block duration
4087          * @see LanguageFi.php for example implementation
4088          */
4089         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4090                 $duration = SpecialBlock::getSuggestedDurations( $this );
4091                 foreach ( $duration as $show => $value ) {
4092                         if ( strcmp( $str, $value ) == 0 ) {
4093                                 return htmlspecialchars( trim( $show ) );
4094                         }
4095                 }
4096
4097                 if ( wfIsInfinity( $str ) ) {
4098                         foreach ( $duration as $show => $value ) {
4099                                 if ( wfIsInfinity( $value ) ) {
4100                                         return htmlspecialchars( trim( $show ) );
4101                                 }
4102                         }
4103                 }
4104
4105                 // If all else fails, return a standard duration or timestamp description.
4106                 $time = strtotime( $str, $now );
4107                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4108                         return $str;
4109                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4110                         // The result differs based on current time, so the difference
4111                         // is a fixed duration length.
4112                         return $this->formatDuration( $time - $now );
4113                 } else { // It's an absolute timestamp.
4114                         if ( $time === 0 ) {
4115                                 // wfTimestamp() handles 0 as current time instead of epoch.
4116                                 $time = '19700101000000';
4117                         }
4118                         if ( $user ) {
4119                                 return $this->userTimeAndDate( $time, $user );
4120                         }
4121                         return $this->timeanddate( $time );
4122                 }
4123         }
4124
4125         /**
4126          * languages like Chinese need to be segmented in order for the diff
4127          * to be of any use
4128          *
4129          * @param string $text
4130          * @return string
4131          */
4132         public function segmentForDiff( $text ) {
4133                 return $text;
4134         }
4135
4136         /**
4137          * and unsegment to show the result
4138          *
4139          * @param string $text
4140          * @return string
4141          */
4142         public function unsegmentForDiff( $text ) {
4143                 return $text;
4144         }
4145
4146         /**
4147          * Return the LanguageConverter used in the Language
4148          *
4149          * @since 1.19
4150          * @return LanguageConverter
4151          */
4152         public function getConverter() {
4153                 return $this->mConverter;
4154         }
4155
4156         /**
4157          * convert text to a variant
4158          *
4159          * @param string $text text to convert
4160          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4161          *      variant (if logged in), or the project default variant
4162          * @return string the converted string
4163          */
4164         public function autoConvert( $text, $variant = false ) {
4165                 return $this->mConverter->autoConvert( $text, $variant );
4166         }
4167
4168         /**
4169          * convert text to all supported variants
4170          *
4171          * @param string $text
4172          * @return array
4173          */
4174         public function autoConvertToAllVariants( $text ) {
4175                 return $this->mConverter->autoConvertToAllVariants( $text );
4176         }
4177
4178         /**
4179          * convert text to different variants of a language.
4180          *
4181          * @warning Glossary state is maintained between calls. This means
4182          *  if you pass unescaped text to this method it can cause an XSS
4183          *  in later calls to this method, even if the later calls have properly
4184          *  escaped the input. Never feed this method user controlled text that
4185          *  is not properly escaped!
4186          * @param string $text Content that has been already escaped for use in HTML
4187          * @return string HTML
4188          */
4189         public function convert( $text ) {
4190                 return $this->mConverter->convert( $text );
4191         }
4192
4193         /**
4194          * Convert a Title object to a string in the preferred variant
4195          *
4196          * @param Title $title
4197          * @return string
4198          */
4199         public function convertTitle( $title ) {
4200                 return $this->mConverter->convertTitle( $title );
4201         }
4202
4203         /**
4204          * Convert a namespace index to a string in the preferred variant
4205          *
4206          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4207          * @param string|null $variant variant to convert to, or null to use the user's preferred
4208          *      variant (if logged in), or the project default variant
4209          * @return string a string representation of the namespace
4210          */
4211         public function convertNamespace( $ns, $variant = null ) {
4212                 return $this->mConverter->convertNamespace( $ns, $variant );
4213         }
4214
4215         /**
4216          * Check if this is a language with variants
4217          *
4218          * @return bool
4219          */
4220         public function hasVariants() {
4221                 return count( $this->getVariants() ) > 1;
4222         }
4223
4224         /**
4225          * Strict check if the language has the specific variant.
4226          *
4227          * Compare to LanguageConverter::validateVariant() which does a more
4228          * lenient check and attempts to coerce the given code to a valid one.
4229          *
4230          * @since 1.19
4231          * @param string $variant
4232          * @return bool
4233          */
4234         public function hasVariant( $variant ) {
4235                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4236         }
4237
4238         /**
4239          * Perform output conversion on a string, and encode for safe HTML output.
4240          * @param string $text Text to be converted
4241          * @return string
4242          * @todo this should get integrated somewhere sane
4243          */
4244         public function convertHtml( $text ) {
4245                 return htmlspecialchars( $this->convert( $text ) );
4246         }
4247
4248         /**
4249          * @param string $key
4250          * @return string
4251          */
4252         public function convertCategoryKey( $key ) {
4253                 return $this->mConverter->convertCategoryKey( $key );
4254         }
4255
4256         /**
4257          * Get the list of variants supported by this language
4258          * see sample implementation in LanguageZh.php
4259          *
4260          * @return string[] An array of language codes
4261          */
4262         public function getVariants() {
4263                 return $this->mConverter->getVariants();
4264         }
4265
4266         /**
4267          * @return string
4268          */
4269         public function getPreferredVariant() {
4270                 return $this->mConverter->getPreferredVariant();
4271         }
4272
4273         /**
4274          * @return string
4275          */
4276         public function getDefaultVariant() {
4277                 return $this->mConverter->getDefaultVariant();
4278         }
4279
4280         /**
4281          * @return string
4282          */
4283         public function getURLVariant() {
4284                 return $this->mConverter->getURLVariant();
4285         }
4286
4287         /**
4288          * If a language supports multiple variants, it is
4289          * possible that non-existing link in one variant
4290          * actually exists in another variant. this function
4291          * tries to find it. See e.g. LanguageZh.php
4292          * The input parameters may be modified upon return
4293          *
4294          * @param string &$link The name of the link
4295          * @param Title &$nt The title object of the link
4296          * @param bool $ignoreOtherCond To disable other conditions when
4297          *   we need to transclude a template or update a category's link
4298          */
4299         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4300                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4301         }
4302
4303         /**
4304          * returns language specific options used by User::getPageRenderHash()
4305          * for example, the preferred language variant
4306          *
4307          * @return string
4308          */
4309         function getExtraHashOptions() {
4310                 return $this->mConverter->getExtraHashOptions();
4311         }
4312
4313         /**
4314          * For languages that support multiple variants, the title of an
4315          * article may be displayed differently in different variants. this
4316          * function returns the apporiate title defined in the body of the article.
4317          *
4318          * @return string
4319          */
4320         public function getParsedTitle() {
4321                 return $this->mConverter->getParsedTitle();
4322         }
4323
4324         /**
4325          * Refresh the cache of conversion tables when
4326          * MediaWiki:Conversiontable* is updated.
4327          *
4328          * @param Title $title The Title of the page being updated
4329          */
4330         public function updateConversionTable( Title $title ) {
4331                 $this->mConverter->updateConversionTable( $title );
4332         }
4333
4334         /**
4335          * Prepare external link text for conversion. When the text is
4336          * a URL, it shouldn't be converted, and it'll be wrapped in
4337          * the "raw" tag (-{R| }-) to prevent conversion.
4338          *
4339          * This function is called "markNoConversion" for historical
4340          * reasons *BUT DIFFERS SIGNIFICANTLY* from
4341          * LanguageConverter::markNoConversion(), with which it is easily
4342          * confused.
4343          *
4344          * @param string $text Text to be used for external link
4345          * @param bool $noParse Wrap it without confirming it's a real URL first
4346          * @return string The tagged text
4347          * @deprecated since 1.32, use LanguageConverter::markNoConversion()
4348          *  instead.
4349          */
4350         public function markNoConversion( $text, $noParse = false ) {
4351                 wfDeprecated( __METHOD__, '1.32' );
4352                 // Excluding protocal-relative URLs may avoid many false positives.
4353                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4354                         return $this->mConverter->markNoConversion( $text );
4355                 } else {
4356                         return $text;
4357                 }
4358         }
4359
4360         /**
4361          * A regular expression to match legal word-trailing characters
4362          * which should be merged onto a link of the form [[foo]]bar.
4363          *
4364          * @return string
4365          */
4366         public function linkTrail() {
4367                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4368         }
4369
4370         /**
4371          * A regular expression character set to match legal word-prefixing
4372          * characters which should be merged onto a link of the form foo[[bar]].
4373          *
4374          * @return string
4375          */
4376         public function linkPrefixCharset() {
4377                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4378         }
4379
4380         /**
4381          * Get the "parent" language which has a converter to convert a "compatible" language
4382          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4383          *
4384          * @return Language|null
4385          * @since 1.22
4386          */
4387         public function getParentLanguage() {
4388                 if ( $this->mParentLanguage !== false ) {
4389                         return $this->mParentLanguage;
4390                 }
4391
4392                 $code = explode( '-', $this->getCode() )[0];
4393                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4394                         $this->mParentLanguage = null;
4395                         return null;
4396                 }
4397                 $lang = self::factory( $code );
4398                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4399                         $this->mParentLanguage = null;
4400                         return null;
4401                 }
4402
4403                 $this->mParentLanguage = $lang;
4404                 return $lang;
4405         }
4406
4407         /**
4408          * Compare with an other language object
4409          *
4410          * @since 1.28
4411          * @param Language $lang
4412          * @return bool
4413          */
4414         public function equals( Language $lang ) {
4415                 return $lang === $this || $lang->getCode() === $this->mCode;
4416         }
4417
4418         /**
4419          * Get the internal language code for this language object
4420          *
4421          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4422          * htmlspecialchars() or similar
4423          *
4424          * @return string
4425          */
4426         public function getCode() {
4427                 return $this->mCode;
4428         }
4429
4430         /**
4431          * Get the code in BCP 47 format which we can use
4432          * inside of html lang="" tags.
4433          *
4434          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4435          * htmlspecialchars() or similar.
4436          *
4437          * @since 1.19
4438          * @return string
4439          */
4440         public function getHtmlCode() {
4441                 if ( is_null( $this->mHtmlCode ) ) {
4442                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4443                 }
4444                 return $this->mHtmlCode;
4445         }
4446
4447         /**
4448          * @param string $code
4449          * @deprecated since 1.32, use Language::factory to create a new object instead.
4450          */
4451         public function setCode( $code ) {
4452                 $this->mCode = $code;
4453                 // Ensure we don't leave incorrect cached data lying around
4454                 $this->mHtmlCode = null;
4455                 $this->mParentLanguage = false;
4456         }
4457
4458         /**
4459          * Get the language code from a file name. Inverse of getFileName()
4460          * @param string $filename $prefix . $languageCode . $suffix
4461          * @param string $prefix Prefix before the language code
4462          * @param string $suffix Suffix after the language code
4463          * @return string Language code, or false if $prefix or $suffix isn't found
4464          */
4465         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4466                 $m = null;
4467                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4468                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4469                 if ( !count( $m ) ) {
4470                         return false;
4471                 }
4472                 return str_replace( '_', '-', strtolower( $m[1] ) );
4473         }
4474
4475         /**
4476          * @param string $code
4477          * @param bool $fallback Whether we're going through language fallback chain
4478          * @return string Name of the language class
4479          */
4480         public static function classFromCode( $code, $fallback = true ) {
4481                 if ( $fallback && $code == 'en' ) {
4482                         return 'Language';
4483                 } else {
4484                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4485                 }
4486         }
4487
4488         /**
4489          * Get the name of a file for a certain language code
4490          * @param string $prefix Prepend this to the filename
4491          * @param string $code Language code
4492          * @param string $suffix Append this to the filename
4493          * @throws MWException
4494          * @return string $prefix . $mangledCode . $suffix
4495          */
4496         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4497                 if ( !self::isValidBuiltInCode( $code ) ) {
4498                         throw new MWException( "Invalid language code \"$code\"" );
4499                 }
4500
4501                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4502         }
4503
4504         /**
4505          * @param string $code
4506          * @return string
4507          */
4508         public static function getMessagesFileName( $code ) {
4509                 global $IP;
4510                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4511                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4512                 return $file;
4513         }
4514
4515         /**
4516          * @param string $code
4517          * @return string
4518          * @throws MWException
4519          * @since 1.23
4520          */
4521         public static function getJsonMessagesFileName( $code ) {
4522                 global $IP;
4523
4524                 if ( !self::isValidBuiltInCode( $code ) ) {
4525                         throw new MWException( "Invalid language code \"$code\"" );
4526                 }
4527
4528                 return "$IP/languages/i18n/$code.json";
4529         }
4530
4531         /**
4532          * Get the first fallback for a given language.
4533          *
4534          * @param string $code
4535          *
4536          * @return bool|string
4537          */
4538         public static function getFallbackFor( $code ) {
4539                 $fallbacks = self::getFallbacksFor( $code );
4540                 if ( $fallbacks ) {
4541                         return $fallbacks[0];
4542                 }
4543                 return false;
4544         }
4545
4546         /**
4547          * Get the ordered list of fallback languages.
4548          *
4549          * @since 1.19
4550          * @param string $code Language code
4551          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4552          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4553          * @throws MWException
4554          * @return array List of language codes
4555          */
4556         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4557                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4558                         return [];
4559                 }
4560                 switch ( $mode ) {
4561                         case self::MESSAGES_FALLBACKS:
4562                                 // For unknown languages, fallbackSequence returns an empty array,
4563                                 // hardcode fallback to 'en' in that case as English messages are
4564                                 // always defined.
4565                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4566                         case self::STRICT_FALLBACKS:
4567                                 // Use this mode when you don't want to fallback to English unless
4568                                 // explicitly defined, for example when you have language-variant icons
4569                                 // and an international language-independent fallback.
4570                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4571                         default:
4572                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4573                 }
4574         }
4575
4576         /**
4577          * Get the ordered list of fallback languages, ending with the fallback
4578          * language chain for the site language.
4579          *
4580          * @since 1.22
4581          * @param string $code Language code
4582          * @return array Array( fallbacks, site fallbacks )
4583          */
4584         public static function getFallbacksIncludingSiteLanguage( $code ) {
4585                 global $wgLanguageCode;
4586
4587                 // Usually, we will only store a tiny number of fallback chains, so we
4588                 // keep them in static memory.
4589                 $cacheKey = "{$code}-{$wgLanguageCode}";
4590
4591                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4592                         $fallbacks = self::getFallbacksFor( $code );
4593
4594                         // Append the site's fallback chain, including the site language itself
4595                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4596                         array_unshift( $siteFallbacks, $wgLanguageCode );
4597
4598                         // Eliminate any languages already included in the chain
4599                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4600
4601                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4602                 }
4603                 return self::$fallbackLanguageCache[$cacheKey];
4604         }
4605
4606         /**
4607          * Get all messages for a given language
4608          * WARNING: this may take a long time. If you just need all message *keys*
4609          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4610          *
4611          * @param string $code
4612          *
4613          * @return array
4614          */
4615         public static function getMessagesFor( $code ) {
4616                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4617         }
4618
4619         /**
4620          * Get a message for a given language
4621          *
4622          * @param string $key
4623          * @param string $code
4624          *
4625          * @return string
4626          */
4627         public static function getMessageFor( $key, $code ) {
4628                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4629         }
4630
4631         /**
4632          * Get all message keys for a given language. This is a faster alternative to
4633          * array_keys( Language::getMessagesFor( $code ) )
4634          *
4635          * @since 1.19
4636          * @param string $code Language code
4637          * @return array Array of message keys (strings)
4638          */
4639         public static function getMessageKeysFor( $code ) {
4640                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4641         }
4642
4643         /**
4644          * @param string $talk
4645          * @return string
4646          */
4647         function fixVariableInNamespace( $talk ) {
4648                 if ( strpos( $talk, '$1' ) === false ) {
4649                         return $talk;
4650                 }
4651
4652                 global $wgMetaNamespace;
4653                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4654
4655                 # Allow grammar transformations
4656                 # Allowing full message-style parsing would make simple requests
4657                 # such as action=raw much more expensive than they need to be.
4658                 # This will hopefully cover most cases.
4659                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4660                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4661                 return str_replace( ' ', '_', $talk );
4662         }
4663
4664         /**
4665          * @param string $m
4666          * @return string
4667          */
4668         function replaceGrammarInNamespace( $m ) {
4669                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4670         }
4671
4672         /**
4673          * Decode an expiry (block, protection, etc) which has come from the DB
4674          *
4675          * @param string $expiry Database expiry String
4676          * @param bool|int $format True to process using language functions, or TS_ constant
4677          *     to return the expiry in a given timestamp
4678          * @param string $infinity If $format is not true, use this string for infinite expiry
4679          * @return string
4680          * @since 1.18
4681          */
4682         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4683                 static $dbInfinity;
4684                 if ( $dbInfinity === null ) {
4685                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4686                 }
4687
4688                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4689                         return $format === true
4690                                 ? $this->getMessageFromDB( 'infiniteblock' )
4691                                 : $infinity;
4692                 } else {
4693                         return $format === true
4694                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4695                                 : wfTimestamp( $format, $expiry );
4696                 }
4697         }
4698
4699         /**
4700          * Formats a time given in seconds into a string representation of that time.
4701          *
4702          * @param int|float $seconds
4703          * @param array $format An optional argument that formats the returned string in different ways:
4704          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4705          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4706          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4707          *     and friends.
4708          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4709          *     or 'avoidminutes'.
4710          * @return string
4711          */
4712         function formatTimePeriod( $seconds, $format = [] ) {
4713                 if ( !is_array( $format ) ) {
4714                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4715                 }
4716                 if ( !isset( $format['avoid'] ) ) {
4717                         $format['avoid'] = false;
4718                 }
4719                 if ( !isset( $format['noabbrevs'] ) ) {
4720                         $format['noabbrevs'] = false;
4721                 }
4722                 $secondsMsg = wfMessage(
4723                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4724                 $minutesMsg = wfMessage(
4725                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4726                 $hoursMsg = wfMessage(
4727                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4728                 $daysMsg = wfMessage(
4729                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4730
4731                 if ( round( $seconds * 10 ) < 100 ) {
4732                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4733                         $s = $secondsMsg->params( $s )->text();
4734                 } elseif ( round( $seconds ) < 60 ) {
4735                         $s = $this->formatNum( round( $seconds ) );
4736                         $s = $secondsMsg->params( $s )->text();
4737                 } elseif ( round( $seconds ) < 3600 ) {
4738                         $minutes = floor( $seconds / 60 );
4739                         $secondsPart = round( fmod( $seconds, 60 ) );
4740                         if ( $secondsPart == 60 ) {
4741                                 $secondsPart = 0;
4742                                 $minutes++;
4743                         }
4744                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4745                         $s .= ' ';
4746                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4747                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4748                         $hours = floor( $seconds / 3600 );
4749                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4750                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4751                         if ( $secondsPart == 60 ) {
4752                                 $secondsPart = 0;
4753                                 $minutes++;
4754                         }
4755                         if ( $minutes == 60 ) {
4756                                 $minutes = 0;
4757                                 $hours++;
4758                         }
4759                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4760                         $s .= ' ';
4761                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4762                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4763                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4764                         }
4765                 } else {
4766                         $days = floor( $seconds / 86400 );
4767                         if ( $format['avoid'] === 'avoidminutes' ) {
4768                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4769                                 if ( $hours == 24 ) {
4770                                         $hours = 0;
4771                                         $days++;
4772                                 }
4773                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4774                                 $s .= ' ';
4775                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4776                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4777                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4778                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4779                                 if ( $minutes == 60 ) {
4780                                         $minutes = 0;
4781                                         $hours++;
4782                                 }
4783                                 if ( $hours == 24 ) {
4784                                         $hours = 0;
4785                                         $days++;
4786                                 }
4787                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4788                                 $s .= ' ';
4789                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4790                                 $s .= ' ';
4791                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4792                         } else {
4793                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4794                                 $s .= ' ';
4795                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4796                         }
4797                 }
4798                 return $s;
4799         }
4800
4801         /**
4802          * Format a bitrate for output, using an appropriate
4803          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4804          *   the magnitude in question.
4805          *
4806          * This use base 1000. For base 1024 use formatSize(), for another base
4807          * see formatComputingNumbers().
4808          *
4809          * @param int $bps
4810          * @return string
4811          */
4812         function formatBitrate( $bps ) {
4813                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4814         }
4815
4816         /**
4817          * @param int $size Size of the unit
4818          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4819          * @param string $messageKey Message key to be uesd
4820          * @return string
4821          */
4822         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4823                 if ( $size <= 0 ) {
4824                         return str_replace( '$1', $this->formatNum( $size ),
4825                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4826                         );
4827                 }
4828                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4829                 $index = 0;
4830
4831                 $maxIndex = count( $sizes ) - 1;
4832                 while ( $size >= $boundary && $index < $maxIndex ) {
4833                         $index++;
4834                         $size /= $boundary;
4835                 }
4836
4837                 // For small sizes no decimal places necessary
4838                 $round = 0;
4839                 if ( $index > 1 ) {
4840                         // For MB and bigger two decimal places are smarter
4841                         $round = 2;
4842                 }
4843                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4844
4845                 $size = round( $size, $round );
4846                 $text = $this->getMessageFromDB( $msg );
4847                 return str_replace( '$1', $this->formatNum( $size ), $text );
4848         }
4849
4850         /**
4851          * Format a size in bytes for output, using an appropriate
4852          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4853          *
4854          * This method use base 1024. For base 1000 use formatBitrate(), for
4855          * another base see formatComputingNumbers()
4856          *
4857          * @param int $size Size to format
4858          * @return string Plain text (not HTML)
4859          */
4860         function formatSize( $size ) {
4861                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4862         }
4863
4864         /**
4865          * Make a list item, used by various special pages
4866          *
4867          * @param string $page Page link
4868          * @param string $details HTML safe text between brackets
4869          * @param bool $oppositedm Add the direction mark opposite to your
4870          *   language, to display text properly
4871          * @return string HTML escaped
4872          */
4873         function specialList( $page, $details, $oppositedm = true ) {
4874                 if ( !$details ) {
4875                         return $page;
4876                 }
4877
4878                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4879                 return $page .
4880                         $dirmark .
4881                         $this->msg( 'word-separator' )->escaped() .
4882                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4883         }
4884
4885         /**
4886          * Generate (prev x| next x) (20|50|100...) type links for paging
4887          *
4888          * @param Title $title Title object to link
4889          * @param int $offset
4890          * @param int $limit
4891          * @param array $query Optional URL query parameter string
4892          * @param bool $atend Optional param for specified if this is the last page
4893          * @return string
4894          */
4895         public function viewPrevNext( Title $title, $offset, $limit,
4896                 array $query = [], $atend = false
4897         ) {
4898                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4899
4900                 # Make 'previous' link
4901                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4902                 if ( $offset > 0 ) {
4903                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4904                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4905                 } else {
4906                         $plink = htmlspecialchars( $prev );
4907                 }
4908
4909                 # Make 'next' link
4910                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4911                 if ( $atend ) {
4912                         $nlink = htmlspecialchars( $next );
4913                 } else {
4914                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4915                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4916                 }
4917
4918                 # Make links to set number of items per page
4919                 $numLinks = [];
4920                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4921                         $numLinks[] = $this->numLink( $title, $offset, $num,
4922                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4923                 }
4924
4925                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4926                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4927         }
4928
4929         /**
4930          * Helper function for viewPrevNext() that generates links
4931          *
4932          * @param Title $title Title object to link
4933          * @param int $offset
4934          * @param int $limit
4935          * @param array $query Extra query parameters
4936          * @param string $link Text to use for the link; will be escaped
4937          * @param string $tooltipMsg Name of the message to use as tooltip
4938          * @param string $class Value of the "class" attribute of the link
4939          * @return string HTML fragment
4940          */
4941         private function numLink( Title $title, $offset, $limit, array $query, $link,
4942                 $tooltipMsg, $class
4943         ) {
4944                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4945                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4946                         ->numParams( $limit )->text();
4947
4948                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4949                         'title' => $tooltip, 'class' => $class ], $link );
4950         }
4951
4952         /**
4953          * Get the conversion rule title, if any.
4954          *
4955          * @return string
4956          */
4957         public function getConvRuleTitle() {
4958                 return $this->mConverter->getConvRuleTitle();
4959         }
4960
4961         /**
4962          * Get the compiled plural rules for the language
4963          * @since 1.20
4964          * @return array Associative array with plural form, and plural rule as key-value pairs
4965          */
4966         public function getCompiledPluralRules() {
4967                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4968                 $fallbacks = self::getFallbacksFor( $this->mCode );
4969                 if ( !$pluralRules ) {
4970                         foreach ( $fallbacks as $fallbackCode ) {
4971                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4972                                 if ( $pluralRules ) {
4973                                         break;
4974                                 }
4975                         }
4976                 }
4977                 return $pluralRules;
4978         }
4979
4980         /**
4981          * Get the plural rules for the language
4982          * @since 1.20
4983          * @return array Associative array with plural form number and plural rule as key-value pairs
4984          */
4985         public function getPluralRules() {
4986                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4987                 $fallbacks = self::getFallbacksFor( $this->mCode );
4988                 if ( !$pluralRules ) {
4989                         foreach ( $fallbacks as $fallbackCode ) {
4990                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4991                                 if ( $pluralRules ) {
4992                                         break;
4993                                 }
4994                         }
4995                 }
4996                 return $pluralRules;
4997         }
4998
4999         /**
5000          * Get the plural rule types for the language
5001          * @since 1.22
5002          * @return array Associative array with plural form number and plural rule type as key-value pairs
5003          */
5004         public function getPluralRuleTypes() {
5005                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
5006                 $fallbacks = self::getFallbacksFor( $this->mCode );
5007                 if ( !$pluralRuleTypes ) {
5008                         foreach ( $fallbacks as $fallbackCode ) {
5009                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
5010                                 if ( $pluralRuleTypes ) {
5011                                         break;
5012                                 }
5013                         }
5014                 }
5015                 return $pluralRuleTypes;
5016         }
5017
5018         /**
5019          * Find the index number of the plural rule appropriate for the given number
5020          * @param int $number
5021          * @return int The index number of the plural rule
5022          */
5023         public function getPluralRuleIndexNumber( $number ) {
5024                 $pluralRules = $this->getCompiledPluralRules();
5025                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5026                 return $form;
5027         }
5028
5029         /**
5030          * Find the plural rule type appropriate for the given number
5031          * For example, if the language is set to Arabic, getPluralType(5) should
5032          * return 'few'.
5033          * @since 1.22
5034          * @param int $number
5035          * @return string The name of the plural rule type, e.g. one, two, few, many
5036          */
5037         public function getPluralRuleType( $number ) {
5038                 $index = $this->getPluralRuleIndexNumber( $number );
5039                 $pluralRuleTypes = $this->getPluralRuleTypes();
5040                 return $pluralRuleTypes[$index] ?? 'other';
5041         }
5042 }