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 array
 511          */
 512         public function getNamespaces() {
 513                 if ( is_null( $this->namespaceNames ) ) {
 514                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 515
 516                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 517
 518                         $this->namespaceNames = $wgExtraNamespaces +
 519                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 520                         $this->namespaceNames += $validNamespaces;
 521
 522                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 523                         if ( $wgMetaNamespaceTalk ) {
 524                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 525                         } else {
 526                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 527                                 $this->namespaceNames[NS_PROJECT_TALK] =
 528                                         $this->fixVariableInNamespace( $talk );
 529                         }
 530
 531                         # Sometimes a language will be localised but not actually exist on this wiki.
 532                         foreach ( $this->namespaceNames as $key => $text ) {
 533                                 if ( !isset( $validNamespaces[$key] ) ) {
 534                                         unset( $this->namespaceNames[$key] );
 535                                 }
 536                         }
 537
 538                         # The above mixing may leave namespaces out of canonical order.
 539                         # Re-order by namespace ID number...
 540                         ksort( $this->namespaceNames );
 541
 542                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 543                 }
 544
 545                 return $this->namespaceNames;
 546         }
 547
 548         /**
 549          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 550          * @param array $namespaces Array of namespaces (id => name)
 551          */
 552         public function setNamespaces( array $namespaces ) {
 553                 $this->namespaceNames = $namespaces;
 554                 $this->mNamespaceIds = null;
 555         }
 556
 557         /**
 558          * Resets all of the namespace caches. Mainly used for testing
 559          */
 560         public function resetNamespaces() {
 561                 $this->namespaceNames = null;
 562                 $this->mNamespaceIds = null;
 563                 $this->namespaceAliases = null;
 564         }
 565
 566         /**
 567          * A convenience function that returns getNamespaces() with spaces instead of underscores
 568          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 569          *
 570          * @return array
 571          */
 572         public function getFormattedNamespaces() {
 573                 $ns = $this->getNamespaces();
 574                 foreach ( $ns as $k => $v ) {
 575                         $ns[$k] = strtr( $v, '_', ' ' );
 576                 }
 577                 return $ns;
 578         }
 579
 580         /**
 581          * Get a namespace value by key
 582          *
 583          * <code>
 584          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 585          * echo $mw_ns; // prints 'MediaWiki'
 586          * </code>
 587          *
 588          * @param int $index The array key of the namespace to return
 589          * @return string|bool String if the namespace value exists, otherwise false
 590          */
 591         public function getNsText( $index ) {
 592                 $ns = $this->getNamespaces();
 593                 return $ns[$index] ?? false;
 594         }
 595
 596         /**
 597          * A convenience function that returns the same thing as
 598          * getNsText() except with '_' changed to ' ', useful for
 599          * producing output.
 600          *
 601          * <code>
 602          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 603          * echo $mw_ns; // prints 'MediaWiki talk'
 604          * </code>
 605          *
 606          * @param int $index The array key of the namespace to return
 607          * @return string Namespace name without underscores (empty string if namespace does not exist)
 608          */
 609         public function getFormattedNsText( $index ) {
 610                 $ns = $this->getNsText( $index );
 611                 return strtr( $ns, '_', ' ' );
 612         }
 613
 614         /**
 615          * Returns gender-dependent namespace alias if available.
 616          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 617          * @param int $index Namespace index
 618          * @param string $gender Gender key (male, female... )
 619          * @return string
 620          * @since 1.18
 621          */
 622         public function getGenderNsText( $index, $gender ) {
 623                 global $wgExtraGenderNamespaces;
 624
 625                 $ns = $wgExtraGenderNamespaces +
 626                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 627
 628                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 629         }
 630
 631         /**
 632          * Whether this language uses gender-dependent namespace aliases.
 633          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 634          * @return bool
 635          * @since 1.18
 636          */
 637         public function needsGenderDistinction() {
 638                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 639                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 640                         // $wgExtraGenderNamespaces overrides everything
 641                         return true;
 642                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 643                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 644                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 645                         return false;
 646                 } else {
 647                         // Check what is in i18n files
 648                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 649                         return count( $aliases ) > 0;
 650                 }
 651         }
 652
 653         /**
 654          * Get a namespace key by value, case insensitive.
 655          * Only matches namespace names for the current language, not the
 656          * canonical ones defined in Namespace.php.
 657          *
 658          * @param string $text
 659          * @return int|bool An integer if $text is a valid value otherwise false
 660          */
 661         function getLocalNsIndex( $text ) {
 662                 $lctext = $this->lc( $text );
 663                 $ids = $this->getNamespaceIds();
 664                 return $ids[$lctext] ?? false;
 665         }
 666
 667         /**
 668          * @return array
 669          */
 670         public function getNamespaceAliases() {
 671                 if ( is_null( $this->namespaceAliases ) ) {
 672                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 673                         if ( !$aliases ) {
 674                                 $aliases = [];
 675                         } else {
 676                                 foreach ( $aliases as $name => $index ) {
 677                                         if ( $index === NS_PROJECT_TALK ) {
 678                                                 unset( $aliases[$name] );
 679                                                 $name = $this->fixVariableInNamespace( $name );
 680                                                 $aliases[$name] = $index;
 681                                         }
 682                                 }
 683                         }
 684
 685                         global $wgExtraGenderNamespaces;
 686                         $genders = $wgExtraGenderNamespaces +
 687                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 688                         foreach ( $genders as $index => $forms ) {
 689                                 foreach ( $forms as $alias ) {
 690                                         $aliases[$alias] = $index;
 691                                 }
 692                         }
 693
 694                         # Also add converted namespace names as aliases, to avoid confusion.
 695                         $convertedNames = [];
 696                         foreach ( $this->getVariants() as $variant ) {
 697                                 if ( $variant === $this->mCode ) {
 698                                         continue;
 699                                 }
 700                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 701                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 702                                 }
 703                         }
 704
 705                         $this->namespaceAliases = $aliases + $convertedNames;
 706
 707                         # 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 ? $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 ? $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                                 # Adar in a leap year
1837                                 if ( $isLeap ) {
1838                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1839                                         $days = 30;
1840                                         if ( $hebrewDay <= $days ) {
1841                                                 # Day in Adar I
1842                                                 $hebrewMonth = 13;
1843                                         } else {
1844                                                 # Subtract the days of Adar I
1845                                                 $hebrewDay -= $days;
1846                                                 # Try Adar II
1847                                                 $days = 29;
1848                                                 if ( $hebrewDay <= $days ) {
1849                                                         # Day in Adar II
1850                                                         $hebrewMonth = 14;
1851                                                 }
1852                                         }
1853                                 }
1854                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1855                                 # Cheshvan in a complete year (otherwise as the rule below)
1856                                 $days = 30;
1857                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1858                                 # Kislev in an incomplete year (otherwise as the rule below)
1859                                 $days = 29;
1860                         } else {
1861                                 # Odd months have 30 days, even have 29
1862                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1863                         }
1864                         if ( $hebrewDay <= $days ) {
1865                                 # In the current month
1866                                 break;
1867                         } else {
1868                                 # Subtract the days of the current month
1869                                 $hebrewDay -= $days;
1870                                 # Try in the next month
1871                                 $hebrewMonth++;
1872                         }
1873                 }
1874
1875                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1876         }
1877
1878         /**
1879          * This calculates the Hebrew year start, as days since 1 September.
1880          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1881          * Used for Hebrew date.
1882          *
1883          * @param int $year
1884          *
1885          * @return string
1886          */
1887         private static function hebrewYearStart( $year ) {
1888                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1889                 $b = intval( ( $year - 1 ) % 4 );
1890                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1891                 if ( $m < 0 ) {
1892                         $m--;
1893                 }
1894                 $Mar = intval( $m );
1895                 if ( $m < 0 ) {
1896                         $m++;
1897                 }
1898                 $m -= $Mar;
1899
1900                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1901                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1902                         $Mar++;
1903                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1904                         $Mar += 2;
1905                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1906                         $Mar++;
1907                 }
1908
1909                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1910                 return $Mar;
1911         }
1912
1913         /**
1914          * Algorithm to convert Gregorian dates to Thai solar dates,
1915          * Minguo dates or Minguo dates.
1916          *
1917          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1918          *       https://en.wikipedia.org/wiki/Minguo_calendar
1919          *       https://en.wikipedia.org/wiki/Japanese_era_name
1920          *
1921          * @param string $ts 14-character timestamp
1922          * @param string $cName Calender name
1923          * @return array Converted year, month, day
1924          */
1925         private static function tsToYear( $ts, $cName ) {
1926                 $gy = substr( $ts, 0, 4 );
1927                 $gm = substr( $ts, 4, 2 );
1928                 $gd = substr( $ts, 6, 2 );
1929
1930                 if ( !strcmp( $cName, 'thai' ) ) {
1931                         # Thai solar dates
1932                         # Add 543 years to the Gregorian calendar
1933                         # Months and days are identical
1934                         $gy_offset = $gy + 543;
1935                         # fix for dates between 1912 and 1941
1936                         # https://en.wikipedia.org/?oldid=836596673#New_year
1937                         if ( $gy >= 1912 && $gy <= 1940 ) {
1938                                 if ( $gm <= 3 ) {
1939                                         $gy_offset--;
1940                                 }
1941                                 $gm = ( $gm - 3 ) % 12;
1942                         }
1943                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1944                         # Minguo dates
1945                         # Deduct 1911 years from the Gregorian calendar
1946                         # Months and days are identical
1947                         $gy_offset = $gy - 1911;
1948                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1949                         # Nengō dates up to Meiji period
1950                         # Deduct years from the Gregorian calendar
1951                         # depending on the nengo periods
1952                         # Months and days are identical
1953                         if ( ( $gy < 1912 )
1954                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1955                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1956                         ) {
1957                                 # Meiji period
1958                                 $gy_gannen = $gy - 1868 + 1;
1959                                 $gy_offset = $gy_gannen;
1960                                 if ( $gy_gannen == 1 ) {
1961                                         $gy_offset = '元';
1962                                 }
1963                                 $gy_offset = '明治' . $gy_offset;
1964                         } elseif (
1965                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1966                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1967                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1968                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1969                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1970                         ) {
1971                                 # Taishō period
1972                                 $gy_gannen = $gy - 1912 + 1;
1973                                 $gy_offset = $gy_gannen;
1974                                 if ( $gy_gannen == 1 ) {
1975                                         $gy_offset = '元';
1976                                 }
1977                                 $gy_offset = '大正' . $gy_offset;
1978                         } elseif (
1979                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1980                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1981                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1982                         ) {
1983                                 # Shōwa period
1984                                 $gy_gannen = $gy - 1926 + 1;
1985                                 $gy_offset = $gy_gannen;
1986                                 if ( $gy_gannen == 1 ) {
1987                                         $gy_offset = '元';
1988                                 }
1989                                 $gy_offset = '昭和' . $gy_offset;
1990                         } else {
1991                                 # Heisei period
1992                                 $gy_gannen = $gy - 1989 + 1;
1993                                 $gy_offset = $gy_gannen;
1994                                 if ( $gy_gannen == 1 ) {
1995                                         $gy_offset = '元';
1996                                 }
1997                                 $gy_offset = '平成' . $gy_offset;
1998                         }
1999                 } else {
2000                         $gy_offset = $gy;
2001                 }
2002
2003                 return [ $gy_offset, $gm, $gd ];
2004         }
2005
2006         /**
2007          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2008          *
2009          * This is the rule the BIDI algorithm uses to determine the directionality of
2010          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2011          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2012          *
2013          * TODO: Does not handle BIDI control characters inside the text.
2014          * TODO: Does not handle unallocated characters.
2015          *
2016          * @param string $text Text to test
2017          * @return null|string Directionality ('ltr' or 'rtl') or null
2018          */
2019         private static function strongDirFromContent( $text = '' ) {
2020                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2021                         return null;
2022                 }
2023                 if ( $matches[1] === '' ) {
2024                         return 'rtl';
2025                 }
2026                 return 'ltr';
2027         }
2028
2029         /**
2030          * Roman number formatting up to 10000
2031          *
2032          * @param int $num
2033          *
2034          * @return string
2035          */
2036         static function romanNumeral( $num ) {
2037                 static $table = [
2038                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2039                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2040                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2041                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2042                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2043                 ];
2044
2045                 $num = intval( $num );
2046                 if ( $num > 10000 || $num <= 0 ) {
2047                         return $num;
2048                 }
2049
2050                 $s = '';
2051                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2052                         if ( $num >= $pow10 ) {
2053                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2054                         }
2055                         $num = $num % $pow10;
2056                 }
2057                 return $s;
2058         }
2059
2060         /**
2061          * Hebrew Gematria number formatting up to 9999
2062          *
2063          * @param int $num
2064          *
2065          * @return string
2066          */
2067         static function hebrewNumeral( $num ) {
2068                 static $table = [
2069                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2070                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2071                         [ '',
2072                                 [ 'ק' ],
2073                                 [ 'ר' ],
2074                                 [ 'ש' ],
2075                                 [ 'ת' ],
2076                                 [ 'ת', 'ק' ],
2077                                 [ 'ת', 'ר' ],
2078                                 [ 'ת', 'ש' ],
2079                                 [ 'ת', 'ת' ],
2080                                 [ 'ת', 'ת', 'ק' ],
2081                                 [ 'ת', 'ת', 'ר' ],
2082                         ],
2083                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2084                 ];
2085
2086                 $num = intval( $num );
2087                 if ( $num > 9999 || $num <= 0 ) {
2088                         return $num;
2089                 }
2090
2091                 // Round thousands have special notations
2092                 if ( $num === 1000 ) {
2093                         return "א' אלף";
2094                 } elseif ( $num % 1000 === 0 ) {
2095                         return $table[0][$num / 1000] . "' אלפים";
2096                 }
2097
2098                 $letters = [];
2099
2100                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2101                         if ( $num >= $pow10 ) {
2102                                 if ( $num === 15 || $num === 16 ) {
2103                                         $letters[] = $table[0][9];
2104                                         $letters[] = $table[0][$num - 9];
2105                                         $num = 0;
2106                                 } else {
2107                                         $letters = array_merge(
2108                                                 $letters,
2109                                                 (array)$table[$i][intval( $num / $pow10 )]
2110                                         );
2111
2112                                         if ( $pow10 === 1000 ) {
2113                                                 $letters[] = "'";
2114                                         }
2115                                 }
2116                         }
2117
2118                         $num = $num % $pow10;
2119                 }
2120
2121                 $preTransformLength = count( $letters );
2122                 if ( $preTransformLength === 1 ) {
2123                         // Add geresh (single quote) to one-letter numbers
2124                         $letters[] = "'";
2125                 } else {
2126                         $lastIndex = $preTransformLength - 1;
2127                         $letters[$lastIndex] = str_replace(
2128                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2129                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2130                                 $letters[$lastIndex]
2131                         );
2132
2133                         // Add gershayim (double quote) to multiple-letter numbers,
2134                         // but exclude numbers with only one letter after the thousands
2135                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2136                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2137                                 $letters[] = "'";
2138                         } else {
2139                                 array_splice( $letters, -1, 0, '"' );
2140                         }
2141                 }
2142
2143                 return implode( $letters );
2144         }
2145
2146         /**
2147          * Used by date() and time() to adjust the time output.
2148          *
2149          * @param string $ts The time in date('YmdHis') format
2150          * @param mixed $tz Adjust the time by this amount (default false, mean we
2151          *   get user timecorrection setting)
2152          * @return int
2153          */
2154         public function userAdjust( $ts, $tz = false ) {
2155                 global $wgUser, $wgLocalTZoffset;
2156
2157                 if ( $tz === false ) {
2158                         $tz = $wgUser->getOption( 'timecorrection' );
2159                 }
2160
2161                 $data = explode( '|', $tz, 3 );
2162
2163                 if ( $data[0] == 'ZoneInfo' ) {
2164                         try {
2165                                 $userTZ = new DateTimeZone( $data[2] );
2166                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2167                                 $date->setTimezone( $userTZ );
2168                                 return $date->format( 'YmdHis' );
2169                         } catch ( Exception $e ) {
2170                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2171                                 $data[0] = 'Offset';
2172                         }
2173                 }
2174
2175                 if ( $data[0] == 'System' || $tz == '' ) {
2176                         # Global offset in minutes.
2177                         $minDiff = $wgLocalTZoffset;
2178                 } elseif ( $data[0] == 'Offset' ) {
2179                         $minDiff = intval( $data[1] );
2180                 } else {
2181                         $data = explode( ':', $tz );
2182                         if ( count( $data ) == 2 ) {
2183                                 $data[0] = intval( $data[0] );
2184                                 $data[1] = intval( $data[1] );
2185                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2186                                 if ( $data[0] < 0 ) {
2187                                         $minDiff = -$minDiff;
2188                                 }
2189                         } else {
2190                                 $minDiff = intval( $data[0] ) * 60;
2191                         }
2192                 }
2193
2194                 # No difference ? Return time unchanged
2195                 if ( 0 == $minDiff ) {
2196                         return $ts;
2197                 }
2198
2199                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2200                 # Generate an adjusted date; take advantage of the fact that mktime
2201                 # will normalize out-of-range values so we don't have to split $minDiff
2202                 # into hours and minutes.
2203                 $t = mktime( (
2204                         (int)substr( $ts, 8, 2 ) ), # Hours
2205                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2206                         (int)substr( $ts, 12, 2 ), # Seconds
2207                         (int)substr( $ts, 4, 2 ), # Month
2208                         (int)substr( $ts, 6, 2 ), # Day
2209                         (int)substr( $ts, 0, 4 ) ); # Year
2210
2211                 $date = date( 'YmdHis', $t );
2212                 Wikimedia\restoreWarnings();
2213
2214                 return $date;
2215         }
2216
2217         /**
2218          * This is meant to be used by time(), date(), and timeanddate() to get
2219          * the date preference they're supposed to use, it should be used in
2220          * all children.
2221          *
2222          *     function timeanddate([...], $format = true) {
2223          *       $datePreference = $this->dateFormat($format);
2224          *       [...]
2225          *     }
2226          *
2227          * @param int|string|bool $usePrefs If true, the user's preference is used
2228          *   if false, the site/language default is used
2229          *   if int/string, assumed to be a format.
2230          * @return string
2231          */
2232         function dateFormat( $usePrefs = true ) {
2233                 global $wgUser;
2234
2235                 if ( is_bool( $usePrefs ) ) {
2236                         if ( $usePrefs ) {
2237                                 $datePreference = $wgUser->getDatePreference();
2238                         } else {
2239                                 $datePreference = (string)User::getDefaultOption( 'date' );
2240                         }
2241                 } else {
2242                         $datePreference = (string)$usePrefs;
2243                 }
2244
2245                 // return int
2246                 if ( $datePreference == '' ) {
2247                         return 'default';
2248                 }
2249
2250                 return $datePreference;
2251         }
2252
2253         /**
2254          * Get a format string for a given type and preference
2255          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2256          * @param string $pref The format name as it appears in Messages*.php under
2257          *  $datePreferences.
2258          *
2259          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2260          *
2261          * @return string
2262          */
2263         function getDateFormatString( $type, $pref ) {
2264                 $wasDefault = false;
2265                 if ( $pref == 'default' ) {
2266                         $wasDefault = true;
2267                         $pref = $this->getDefaultDateFormat();
2268                 }
2269
2270                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2271                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2272
2273                         if ( $type === 'pretty' && $df === null ) {
2274                                 $df = $this->getDateFormatString( 'date', $pref );
2275                         }
2276
2277                         if ( !$wasDefault && $df === null ) {
2278                                 $pref = $this->getDefaultDateFormat();
2279                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2280                         }
2281
2282                         $this->dateFormatStrings[$type][$pref] = $df;
2283                 }
2284                 return $this->dateFormatStrings[$type][$pref];
2285         }
2286
2287         /**
2288          * @param string $ts The time format which needs to be turned into a
2289          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2290          * @param bool $adj Whether to adjust the time output according to the
2291          *   user configured offset ($timecorrection)
2292          * @param mixed $format True to use user's date format preference
2293          * @param string|bool $timecorrection The time offset as returned by
2294          *   validateTimeZone() in Special:Preferences
2295          * @return string
2296          */
2297         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2298                 $ts = wfTimestamp( TS_MW, $ts );
2299                 if ( $adj ) {
2300                         $ts = $this->userAdjust( $ts, $timecorrection );
2301                 }
2302                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2303                 return $this->sprintfDate( $df, $ts );
2304         }
2305
2306         /**
2307          * @param string $ts The time format which needs to be turned into a
2308          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2309          * @param bool $adj Whether to adjust the time output according to the
2310          *   user configured offset ($timecorrection)
2311          * @param mixed $format True to use user's date format preference
2312          * @param string|bool $timecorrection The time offset as returned by
2313          *   validateTimeZone() in Special:Preferences
2314          * @return string
2315          */
2316         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2317                 $ts = wfTimestamp( TS_MW, $ts );
2318                 if ( $adj ) {
2319                         $ts = $this->userAdjust( $ts, $timecorrection );
2320                 }
2321                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2322                 return $this->sprintfDate( $df, $ts );
2323         }
2324
2325         /**
2326          * @param string $ts The time format which needs to be turned into a
2327          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2328          * @param bool $adj Whether to adjust the time output according to the
2329          *   user configured offset ($timecorrection)
2330          * @param mixed $format What format to return, if it's false output the
2331          *   default one (default true)
2332          * @param string|bool $timecorrection The time offset as returned by
2333          *   validateTimeZone() in Special:Preferences
2334          * @return string
2335          */
2336         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2337                 $ts = wfTimestamp( TS_MW, $ts );
2338                 if ( $adj ) {
2339                         $ts = $this->userAdjust( $ts, $timecorrection );
2340                 }
2341                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2342                 return $this->sprintfDate( $df, $ts );
2343         }
2344
2345         /**
2346          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2347          *
2348          * @since 1.20
2349          *
2350          * @param int $seconds The amount of seconds.
2351          * @param array $chosenIntervals The intervals to enable.
2352          *
2353          * @return string
2354          */
2355         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2356                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2357
2358                 $segments = [];
2359
2360                 foreach ( $intervals as $intervalName => $intervalValue ) {
2361                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2362                         // duration-years, duration-decades, duration-centuries, duration-millennia
2363                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2364                         $segments[] = $message->inLanguage( $this )->escaped();
2365                 }
2366
2367                 return $this->listToText( $segments );
2368         }
2369
2370         /**
2371          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2372          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2373          *
2374          * @since 1.20
2375          *
2376          * @param int $seconds The amount of seconds.
2377          * @param array $chosenIntervals The intervals to enable.
2378          *
2379          * @return array
2380          */
2381         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2382                 if ( empty( $chosenIntervals ) ) {
2383                         $chosenIntervals = [
2384                                 'millennia',
2385                                 'centuries',
2386                                 'decades',
2387                                 'years',
2388                                 'days',
2389                                 'hours',
2390                                 'minutes',
2391                                 'seconds'
2392                         ];
2393                 }
2394
2395                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2396                 $sortedNames = array_keys( $intervals );
2397                 $smallestInterval = array_pop( $sortedNames );
2398
2399                 $segments = [];
2400
2401                 foreach ( $intervals as $name => $length ) {
2402                         $value = floor( $seconds / $length );
2403
2404                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2405                                 $seconds -= $value * $length;
2406                                 $segments[$name] = $value;
2407                         }
2408                 }
2409
2410                 return $segments;
2411         }
2412
2413         /**
2414          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2415          *
2416          * @param string $type Can be 'date', 'time' or 'both'
2417          * @param string $ts The time format which needs to be turned into a
2418          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2419          * @param User $user User object used to get preferences for timezone and format
2420          * @param array $options Array, can contain the following keys:
2421          *   - 'timecorrection': time correction, can have the following values:
2422          *     - true: use user's preference
2423          *     - false: don't use time correction
2424          *     - int: value of time correction in minutes
2425          *   - 'format': format to use, can have the following values:
2426          *     - true: use user's preference
2427          *     - false: use default preference
2428          *     - string: format to use
2429          * @since 1.19
2430          * @return string
2431          */
2432         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2433                 $ts = wfTimestamp( TS_MW, $ts );
2434                 $options += [ 'timecorrection' => true, 'format' => true ];
2435                 if ( $options['timecorrection'] !== false ) {
2436                         if ( $options['timecorrection'] === true ) {
2437                                 $offset = $user->getOption( 'timecorrection' );
2438                         } else {
2439                                 $offset = $options['timecorrection'];
2440                         }
2441                         $ts = $this->userAdjust( $ts, $offset );
2442                 }
2443                 if ( $options['format'] === true ) {
2444                         $format = $user->getDatePreference();
2445                 } else {
2446                         $format = $options['format'];
2447                 }
2448                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2449                 return $this->sprintfDate( $df, $ts );
2450         }
2451
2452         /**
2453          * Get the formatted date for the given timestamp and formatted for
2454          * the given user.
2455          *
2456          * @param mixed $ts Mixed: the time format which needs to be turned into a
2457          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2458          * @param User $user User object used to get preferences for timezone and format
2459          * @param array $options Array, can contain the following keys:
2460          *   - 'timecorrection': time correction, can have the following values:
2461          *     - true: use user's preference
2462          *     - false: don't use time correction
2463          *     - int: value of time correction in minutes
2464          *   - 'format': format to use, can have the following values:
2465          *     - true: use user's preference
2466          *     - false: use default preference
2467          *     - string: format to use
2468          * @since 1.19
2469          * @return string
2470          */
2471         public function userDate( $ts, User $user, array $options = [] ) {
2472                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2473         }
2474
2475         /**
2476          * Get the formatted time for the given timestamp and formatted for
2477          * the given user.
2478          *
2479          * @param mixed $ts The time format which needs to be turned into a
2480          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2481          * @param User $user User object used to get preferences for timezone and format
2482          * @param array $options Array, can contain the following keys:
2483          *   - 'timecorrection': time correction, can have the following values:
2484          *     - true: use user's preference
2485          *     - false: don't use time correction
2486          *     - int: value of time correction in minutes
2487          *   - 'format': format to use, can have the following values:
2488          *     - true: use user's preference
2489          *     - false: use default preference
2490          *     - string: format to use
2491          * @since 1.19
2492          * @return string
2493          */
2494         public function userTime( $ts, User $user, array $options = [] ) {
2495                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2496         }
2497
2498         /**
2499          * Get the formatted date and time for the given timestamp and formatted for
2500          * the given user.
2501          *
2502          * @param mixed $ts The time format which needs to be turned into a
2503          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2504          * @param User $user User object used to get preferences for timezone and format
2505          * @param array $options Array, can contain the following keys:
2506          *   - 'timecorrection': time correction, can have the following values:
2507          *     - true: use user's preference
2508          *     - false: don't use time correction
2509          *     - int: value of time correction in minutes
2510          *   - 'format': format to use, can have the following values:
2511          *     - true: use user's preference
2512          *     - false: use default preference
2513          *     - string: format to use
2514          * @since 1.19
2515          * @return string
2516          */
2517         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2518                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2519         }
2520
2521         /**
2522          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2523          *
2524          * Determine the difference between the timestamp and the current time, and
2525          * generate a readable timestamp by returning "<N> <units> ago", where the
2526          * largest possible unit is used.
2527          *
2528          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2529          *
2530          * @param MWTimestamp $time
2531          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2532          * @param User|null $user User the timestamp is being generated for
2533          *  (or null to use main context's user)
2534          * @return string Formatted timestamp
2535          */
2536         public function getHumanTimestamp(
2537                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2538         ) {
2539                 if ( $relativeTo === null ) {
2540                         $relativeTo = new MWTimestamp();
2541                 }
2542                 if ( $user === null ) {
2543                         $user = RequestContext::getMain()->getUser();
2544                 }
2545
2546                 // Adjust for the user's timezone.
2547                 $offsetThis = $time->offsetForUser( $user );
2548                 $offsetRel = $relativeTo->offsetForUser( $user );
2549
2550                 $ts = '';
2551                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2552                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2553                 }
2554
2555                 // Reset the timezone on the objects.
2556                 $time->timestamp->sub( $offsetThis );
2557                 $relativeTo->timestamp->sub( $offsetRel );
2558
2559                 return $ts;
2560         }
2561
2562         /**
2563          * Convert an MWTimestamp into a pretty human-readable timestamp using
2564          * the given user preferences and relative base time.
2565          *
2566          * @see Language::getHumanTimestamp
2567          * @param MWTimestamp $ts Timestamp to prettify
2568          * @param MWTimestamp $relativeTo Base timestamp
2569          * @param User $user User preferences to use
2570          * @return string Human timestamp
2571          * @since 1.26
2572          */
2573         private function getHumanTimestampInternal(
2574                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2575         ) {
2576                 $diff = $ts->diff( $relativeTo );
2577                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2578                         (int)$relativeTo->timestamp->format( 'w' ) );
2579                 $days = $diff->days ?: (int)$diffDay;
2580                 if ( $diff->invert || $days > 5
2581                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2582                 ) {
2583                         // Timestamps are in different years: use full timestamp
2584                         // Also do full timestamp for future dates
2585                         /**
2586                          * @todo FIXME: Add better handling of future timestamps.
2587                          */
2588                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2589                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2590                 } elseif ( $days > 5 ) {
2591                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2592                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2593                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2594                 } elseif ( $days > 1 ) {
2595                         // Timestamp within the past week: show the day of the week and time
2596                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2597                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2598                         // Messages:
2599                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2600                         $ts = wfMessage( "$weekday-at" )
2601                                 ->inLanguage( $this )
2602                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2603                                 ->text();
2604                 } elseif ( $days == 1 ) {
2605                         // Timestamp was yesterday: say 'yesterday' and the time.
2606                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2607                         $ts = wfMessage( 'yesterday-at' )
2608                                 ->inLanguage( $this )
2609                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2610                                 ->text();
2611                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2612                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2613                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2614                         $ts = wfMessage( 'today-at' )
2615                                 ->inLanguage( $this )
2616                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2617                                 ->text();
2618
2619                 // From here on in, the timestamp was soon enough ago so that we can simply say
2620                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2621                 } elseif ( $diff->h == 1 ) {
2622                         // Less than 90 minutes, but more than an hour ago.
2623                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2624                 } elseif ( $diff->i >= 1 ) {
2625                         // A few minutes ago.
2626                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2627                 } elseif ( $diff->s >= 30 ) {
2628                         // Less than a minute, but more than 30 sec ago.
2629                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2630                 } else {
2631                         // Less than 30 seconds ago.
2632                         $ts = wfMessage( 'just-now' )->text();
2633                 }
2634
2635                 return $ts;
2636         }
2637
2638         /**
2639          * @param string $key
2640          * @return string|null
2641          */
2642         public function getMessage( $key ) {
2643                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2644         }
2645
2646         /**
2647          * @return array
2648          */
2649         function getAllMessages() {
2650                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2651         }
2652
2653         /**
2654          * @param string $in
2655          * @param string $out
2656          * @param string $string
2657          * @return string
2658          */
2659         public function iconv( $in, $out, $string ) {
2660                 # Even with //IGNORE iconv can whine about illegal characters in
2661                 # *input* string. We just ignore those too.
2662                 # REF: https://bugs.php.net/bug.php?id=37166
2663                 # REF: https://phabricator.wikimedia.org/T18885
2664                 Wikimedia\suppressWarnings();
2665                 $text = iconv( $in, $out . '//IGNORE', $string );
2666                 Wikimedia\restoreWarnings();
2667                 return $text;
2668         }
2669
2670         // callback functions for ucwords(), ucwordbreaks()
2671
2672         /**
2673          * @param array $matches
2674          * @return mixed|string
2675          */
2676         function ucwordbreaksCallbackAscii( $matches ) {
2677                 return $this->ucfirst( $matches[1] );
2678         }
2679
2680         /**
2681          * @param array $matches
2682          * @return string
2683          */
2684         function ucwordbreaksCallbackMB( $matches ) {
2685                 return mb_strtoupper( $matches[0] );
2686         }
2687
2688         /**
2689          * @param array $matches
2690          * @return string
2691          */
2692         function ucwordsCallbackMB( $matches ) {
2693                 return mb_strtoupper( $matches[0] );
2694         }
2695
2696         /**
2697          * Make a string's first character uppercase
2698          *
2699          * @param string $str
2700          *
2701          * @return string
2702          */
2703         public function ucfirst( $str ) {
2704                 $o = ord( $str );
2705                 if ( $o < 96 ) { // if already uppercase...
2706                         return $str;
2707                 } elseif ( $o < 128 ) {
2708                         return ucfirst( $str ); // use PHP's ucfirst()
2709                 } else {
2710                         // fall back to more complex logic in case of multibyte strings
2711                         return $this->uc( $str, true );
2712                 }
2713         }
2714
2715         /**
2716          * Convert a string to uppercase
2717          *
2718          * @param string $str
2719          * @param bool $first
2720          *
2721          * @return string
2722          */
2723         public function uc( $str, $first = false ) {
2724                 if ( $first ) {
2725                         if ( $this->isMultibyte( $str ) ) {
2726                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2727                         } else {
2728                                 return ucfirst( $str );
2729                         }
2730                 } else {
2731                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2732                 }
2733         }
2734
2735         /**
2736          * @param string $str
2737          * @return mixed|string
2738          */
2739         function lcfirst( $str ) {
2740                 $o = ord( $str );
2741                 if ( !$o ) {
2742                         return strval( $str );
2743                 } elseif ( $o >= 128 ) {
2744                         return $this->lc( $str, true );
2745                 } elseif ( $o > 96 ) {
2746                         return $str;
2747                 } else {
2748                         $str[0] = strtolower( $str[0] );
2749                         return $str;
2750                 }
2751         }
2752
2753         /**
2754          * @param string $str
2755          * @param bool $first
2756          * @return mixed|string
2757          */
2758         function lc( $str, $first = false ) {
2759                 if ( $first ) {
2760                         if ( $this->isMultibyte( $str ) ) {
2761                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2762                         } else {
2763                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2764                         }
2765                 } else {
2766                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2767                 }
2768         }
2769
2770         /**
2771          * @param string $str
2772          * @return bool
2773          */
2774         function isMultibyte( $str ) {
2775                 return strlen( $str ) !== mb_strlen( $str );
2776         }
2777
2778         /**
2779          * @param string $str
2780          * @return mixed|string
2781          */
2782         function ucwords( $str ) {
2783                 if ( $this->isMultibyte( $str ) ) {
2784                         $str = $this->lc( $str );
2785
2786                         // regexp to find first letter in each word (i.e. after each space)
2787                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2788
2789                         // function to use to capitalize a single char
2790                         return preg_replace_callback(
2791                                 $replaceRegexp,
2792                                 [ $this, 'ucwordsCallbackMB' ],
2793                                 $str
2794                         );
2795                 } else {
2796                         return ucwords( strtolower( $str ) );
2797                 }
2798         }
2799
2800         /**
2801          * capitalize words at word breaks
2802          *
2803          * @param string $str
2804          * @return mixed
2805          */
2806         function ucwordbreaks( $str ) {
2807                 if ( $this->isMultibyte( $str ) ) {
2808                         $str = $this->lc( $str );
2809
2810                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2811                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2812
2813                         // find first letter after word break
2814                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2815                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2816
2817                         return preg_replace_callback(
2818                                 $replaceRegexp,
2819                                 [ $this, 'ucwordbreaksCallbackMB' ],
2820                                 $str
2821                         );
2822                 } else {
2823                         return preg_replace_callback(
2824                                 '/\b([\w\x80-\xff]+)\b/',
2825                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2826                                 $str
2827                         );
2828                 }
2829         }
2830
2831         /**
2832          * Return a case-folded representation of $s
2833          *
2834          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2835          * and $s2 are the same except for the case of their characters. It is not
2836          * necessary for the value returned to make sense when displayed.
2837          *
2838          * Do *not* perform any other normalisation in this function. If a caller
2839          * uses this function when it should be using a more general normalisation
2840          * function, then fix the caller.
2841          *
2842          * @param string $s
2843          *
2844          * @return string
2845          */
2846         function caseFold( $s ) {
2847                 return $this->uc( $s );
2848         }
2849
2850         /**
2851          * @param string $s
2852          * @return string
2853          * @throws MWException
2854          */
2855         function checkTitleEncoding( $s ) {
2856                 if ( is_array( $s ) ) {
2857                         throw new MWException( 'Given array to checkTitleEncoding.' );
2858                 }
2859                 if ( StringUtils::isUtf8( $s ) ) {
2860                         return $s;
2861                 }
2862
2863                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2864         }
2865
2866         /**
2867          * @return array
2868          */
2869         function fallback8bitEncoding() {
2870                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2871         }
2872
2873         /**
2874          * Most writing systems use whitespace to break up words.
2875          * Some languages such as Chinese don't conventionally do this,
2876          * which requires special handling when breaking up words for
2877          * searching etc.
2878          *
2879          * @return bool
2880          */
2881         function hasWordBreaks() {
2882                 return true;
2883         }
2884
2885         /**
2886          * Some languages such as Chinese require word segmentation,
2887          * Specify such segmentation when overridden in derived class.
2888          *
2889          * @param string $string
2890          * @return string
2891          */
2892         function segmentByWord( $string ) {
2893                 return $string;
2894         }
2895
2896         /**
2897          * Some languages have special punctuation need to be normalized.
2898          * Make such changes here.
2899          *
2900          * @param string $string
2901          * @return string
2902          */
2903         function normalizeForSearch( $string ) {
2904                 return self::convertDoubleWidth( $string );
2905         }
2906
2907         /**
2908          * convert double-width roman characters to single-width.
2909          * range: ff00-ff5f ~= 0020-007f
2910          *
2911          * @param string $string
2912          *
2913          * @return string
2914          */
2915         protected static function convertDoubleWidth( $string ) {
2916                 static $full = null;
2917                 static $half = null;
2918
2919                 if ( $full === null ) {
2920                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2921                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2922                         $full = str_split( $fullWidth, 3 );
2923                         $half = str_split( $halfWidth );
2924                 }
2925
2926                 $string = str_replace( $full, $half, $string );
2927                 return $string;
2928         }
2929
2930         /**
2931          * @param string $string
2932          * @param string $pattern
2933          * @return string
2934          */
2935         protected static function insertSpace( $string, $pattern ) {
2936                 $string = preg_replace( $pattern, " $1 ", $string );
2937                 $string = preg_replace( '/ +/', ' ', $string );
2938                 return $string;
2939         }
2940
2941         /**
2942          * @param array $termsArray
2943          * @return array
2944          */
2945         function convertForSearchResult( $termsArray ) {
2946                 # some languages, e.g. Chinese, need to do a conversion
2947                 # in order for search results to be displayed correctly
2948                 return $termsArray;
2949         }
2950
2951         /**
2952          * Get the first character of a string.
2953          *
2954          * @param string $s
2955          * @return string
2956          */
2957         function firstChar( $s ) {
2958                 $matches = [];
2959                 preg_match(
2960                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2961                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2962                         $s,
2963                         $matches
2964                 );
2965
2966                 if ( isset( $matches[1] ) ) {
2967                         if ( strlen( $matches[1] ) != 3 ) {
2968                                 return $matches[1];
2969                         }
2970
2971                         // Break down Hangul syllables to grab the first jamo
2972                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2973                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2974                                 return $matches[1];
2975                         } elseif ( $code < 0xb098 ) {
2976                                 return "\u{3131}";
2977                         } elseif ( $code < 0xb2e4 ) {
2978                                 return "\u{3134}";
2979                         } elseif ( $code < 0xb77c ) {
2980                                 return "\u{3137}";
2981                         } elseif ( $code < 0xb9c8 ) {
2982                                 return "\u{3139}";
2983                         } elseif ( $code < 0xbc14 ) {
2984                                 return "\u{3141}";
2985                         } elseif ( $code < 0xc0ac ) {
2986                                 return "\u{3142}";
2987                         } elseif ( $code < 0xc544 ) {
2988                                 return "\u{3145}";
2989                         } elseif ( $code < 0xc790 ) {
2990                                 return "\u{3147}";
2991                         } elseif ( $code < 0xcc28 ) {
2992                                 return "\u{3148}";
2993                         } elseif ( $code < 0xce74 ) {
2994                                 return "\u{314A}";
2995                         } elseif ( $code < 0xd0c0 ) {
2996                                 return "\u{314B}";
2997                         } elseif ( $code < 0xd30c ) {
2998                                 return "\u{314C}";
2999                         } elseif ( $code < 0xd558 ) {
3000                                 return "\u{314D}";
3001                         } else {
3002                                 return "\u{314E}";
3003                         }
3004                 } else {
3005                         return '';
3006                 }
3007         }
3008
3009         /**
3010          * @deprecated No-op since 1.28
3011          */
3012         function initEncoding() {
3013                 wfDeprecated( __METHOD__, '1.28' );
3014                 // No-op.
3015         }
3016
3017         /**
3018          * @param string $s
3019          * @return string
3020          * @deprecated No-op since 1.28
3021          */
3022         function recodeForEdit( $s ) {
3023                 wfDeprecated( __METHOD__, '1.28' );
3024                 return $s;
3025         }
3026
3027         /**
3028          * @param string $s
3029          * @return string
3030          * @deprecated No-op since 1.28
3031          */
3032         function recodeInput( $s ) {
3033                 wfDeprecated( __METHOD__, '1.28' );
3034                 return $s;
3035         }
3036
3037         /**
3038          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3039          * also cleans up certain backwards-compatible sequences, converting them
3040          * to the modern Unicode equivalent.
3041          *
3042          * This is language-specific for performance reasons only.
3043          *
3044          * @param string $s
3045          *
3046          * @return string
3047          */
3048         public function normalize( $s ) {
3049                 global $wgAllUnicodeFixes;
3050                 $s = UtfNormal\Validator::cleanUp( $s );
3051                 if ( $wgAllUnicodeFixes ) {
3052                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3053                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3054                 }
3055
3056                 return $s;
3057         }
3058
3059         /**
3060          * Transform a string using serialized data stored in the given file (which
3061          * must be in the serialized subdirectory of $IP). The file contains pairs
3062          * mapping source characters to destination characters.
3063          *
3064          * The data is cached in process memory. This will go faster if you have the
3065          * FastStringSearch extension.
3066          *
3067          * @param string $file
3068          * @param string $string
3069          *
3070          * @throws MWException
3071          * @return string
3072          */
3073         protected function transformUsingPairFile( $file, $string ) {
3074                 if ( !isset( $this->transformData[$file] ) ) {
3075                         global $IP;
3076                         $data = require "$IP/languages/data/{$file}";
3077                         $this->transformData[$file] = new ReplacementArray( $data );
3078                 }
3079                 return $this->transformData[$file]->replace( $string );
3080         }
3081
3082         /**
3083          * For right-to-left language support
3084          *
3085          * @return bool
3086          */
3087         function isRTL() {
3088                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3089         }
3090
3091         /**
3092          * Return the correct HTML 'dir' attribute value for this language.
3093          * @return string
3094          */
3095         function getDir() {
3096                 return $this->isRTL() ? 'rtl' : 'ltr';
3097         }
3098
3099         /**
3100          * Return 'left' or 'right' as appropriate alignment for line-start
3101          * for this language's text direction.
3102          *
3103          * Should be equivalent to CSS3 'start' text-align value....
3104          *
3105          * @return string
3106          */
3107         function alignStart() {
3108                 return $this->isRTL() ? 'right' : 'left';
3109         }
3110
3111         /**
3112          * Return 'right' or 'left' as appropriate alignment for line-end
3113          * for this language's text direction.
3114          *
3115          * Should be equivalent to CSS3 'end' text-align value....
3116          *
3117          * @return string
3118          */
3119         function alignEnd() {
3120                 return $this->isRTL() ? 'left' : 'right';
3121         }
3122
3123         /**
3124          * A hidden direction mark (LRM or RLM), depending on the language direction.
3125          * Unlike getDirMark(), this function returns the character as an HTML entity.
3126          * This function should be used when the output is guaranteed to be HTML,
3127          * because it makes the output HTML source code more readable. When
3128          * the output is plain text or can be escaped, getDirMark() should be used.
3129          *
3130          * @param bool $opposite Get the direction mark opposite to your language
3131          * @return string
3132          * @since 1.20
3133          */
3134         function getDirMarkEntity( $opposite = false ) {
3135                 if ( $opposite ) {
3136                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3137                 }
3138                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3139         }
3140
3141         /**
3142          * A hidden direction mark (LRM or RLM), depending on the language direction.
3143          * This function produces them as invisible Unicode characters and
3144          * the output may be hard to read and debug, so it should only be used
3145          * when the output is plain text or can be escaped. When the output is
3146          * HTML, use getDirMarkEntity() instead.
3147          *
3148          * @param bool $opposite Get the direction mark opposite to your language
3149          * @return string
3150          */
3151         function getDirMark( $opposite = false ) {
3152                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3153                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3154                 if ( $opposite ) {
3155                         return $this->isRTL() ? $lrm : $rlm;
3156                 }
3157                 return $this->isRTL() ? $rlm : $lrm;
3158         }
3159
3160         /**
3161          * @return array
3162          */
3163         function capitalizeAllNouns() {
3164                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3165         }
3166
3167         /**
3168          * An arrow, depending on the language direction.
3169          *
3170          * @param string $direction The direction of the arrow: forwards (default),
3171          *   backwards, left, right, up, down.
3172          * @return string
3173          */
3174         function getArrow( $direction = 'forwards' ) {
3175                 switch ( $direction ) {
3176                         case 'forwards':
3177                                 return $this->isRTL() ? '←' : '→';
3178                         case 'backwards':
3179                                 return $this->isRTL() ? '→' : '←';
3180                         case 'left':
3181                                 return '←';
3182                         case 'right':
3183                                 return '→';
3184                         case 'up':
3185                                 return '↑';
3186                         case 'down':
3187                                 return '↓';
3188                 }
3189         }
3190
3191         /**
3192          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3193          *
3194          * @return bool
3195          */
3196         function linkPrefixExtension() {
3197                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3198         }
3199
3200         /**
3201          * Get all magic words from cache.
3202          * @return array
3203          */
3204         function getMagicWords() {
3205                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3206         }
3207
3208         /**
3209          * Fill a MagicWord object with data from here
3210          *
3211          * @param MagicWord $mw
3212          */
3213         function getMagic( $mw ) {
3214                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3215                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3216
3217                 if ( !is_array( $rawEntry ) ) {
3218                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3219                 } else {
3220                         $mw->mCaseSensitive = $rawEntry[0];
3221                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3222                 }
3223         }
3224
3225         /**
3226          * Add magic words to the extension array
3227          *
3228          * @param array $newWords
3229          */
3230         function addMagicWordsByLang( $newWords ) {
3231                 $fallbackChain = $this->getFallbackLanguages();
3232                 $fallbackChain = array_reverse( $fallbackChain );
3233                 foreach ( $fallbackChain as $code ) {
3234                         if ( isset( $newWords[$code] ) ) {
3235                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3236                         }
3237                 }
3238         }
3239
3240         /**
3241          * Get special page names, as an associative array
3242          *   canonical name => array of valid names, including aliases
3243          * @return array
3244          */
3245         function getSpecialPageAliases() {
3246                 // Cache aliases because it may be slow to load them
3247                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3248                         // Initialise array
3249                         $this->mExtendedSpecialPageAliases =
3250                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3251                 }
3252
3253                 return $this->mExtendedSpecialPageAliases;
3254         }
3255
3256         /**
3257          * Italic is unsuitable for some languages
3258          *
3259          * @param string $text The text to be emphasized.
3260          * @return string
3261          */
3262         function emphasize( $text ) {
3263                 return "<em>$text</em>";
3264         }
3265
3266         /**
3267          * Normally we output all numbers in plain en_US style, that is
3268          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3269          * point twohundredthirtyfive. However this is not suitable for all
3270          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3271          * Icelandic just want to use commas instead of dots, and dots instead
3272          * of commas like "293.291,235".
3273          *
3274          * An example of this function being called:
3275          * <code>
3276          * wfMessage( 'message' )->numParams( $num )->text()
3277          * </code>
3278          *
3279          * See $separatorTransformTable on MessageIs.php for
3280          * the , => . and . => , implementation.
3281          *
3282          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3283          * @param int|float $number The string to be formatted, should be an integer
3284          *   or a floating point number.
3285          * @param bool $nocommafy Set to true for special numbers like dates
3286          * @return string
3287          */
3288         public function formatNum( $number, $nocommafy = false ) {
3289                 global $wgTranslateNumerals;
3290                 if ( !$nocommafy ) {
3291                         $number = $this->commafy( $number );
3292                         $s = $this->separatorTransformTable();
3293                         if ( $s ) {
3294                                 $number = strtr( $number, $s );
3295                         }
3296                 }
3297
3298                 if ( $wgTranslateNumerals ) {
3299                         $s = $this->digitTransformTable();
3300                         if ( $s ) {
3301                                 $number = strtr( $number, $s );
3302                         }
3303                 }
3304
3305                 return (string)$number;
3306         }
3307
3308         /**
3309          * Front-end for non-commafied formatNum
3310          *
3311          * @param int|float $number The string to be formatted, should be an integer
3312          *        or a floating point number.
3313          * @since 1.21
3314          * @return string
3315          */
3316         public function formatNumNoSeparators( $number ) {
3317                 return $this->formatNum( $number, true );
3318         }
3319
3320         /**
3321          * @param string $number
3322          * @return string
3323          */
3324         public function parseFormattedNumber( $number ) {
3325                 $s = $this->digitTransformTable();
3326                 if ( $s ) {
3327                         // eliminate empty array values such as ''. (T66347)
3328                         $s = array_filter( $s );
3329                         $number = strtr( $number, array_flip( $s ) );
3330                 }
3331
3332                 $s = $this->separatorTransformTable();
3333                 if ( $s ) {
3334                         // eliminate empty array values such as ''. (T66347)
3335                         $s = array_filter( $s );
3336                         $number = strtr( $number, array_flip( $s ) );
3337                 }
3338
3339                 $number = strtr( $number, [ ',' => '' ] );
3340                 return $number;
3341         }
3342
3343         /**
3344          * Adds commas to a given number
3345          * @since 1.19
3346          * @param mixed $number
3347          * @return string
3348          */
3349         function commafy( $number ) {
3350                 $digitGroupingPattern = $this->digitGroupingPattern();
3351                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3352                 if ( $number === null ) {
3353                         return '';
3354                 }
3355
3356                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3357                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3358                         // In some languages it's conventional not to insert a thousands separator
3359                         // in numbers that are four digits long (1000-9999).
3360                         if ( $minimumGroupingDigits ) {
3361                                 // Number of '#' characters after last comma in the grouping pattern.
3362                                 // The pattern is hardcoded here, but this would vary for different patterns.
3363                                 $primaryGroupingSize = 3;
3364                                 // Maximum length of a number to suppress digit grouping for.
3365                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3366                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3367                                         return $number;
3368                                 }
3369                         }
3370                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3371                 } else {
3372                         // Ref: http://cldr.unicode.org/translation/number-patterns
3373                         $sign = "";
3374                         if ( intval( $number ) < 0 ) {
3375                                 // For negative numbers apply the algorithm like positive number and add sign.
3376                                 $sign = "-";
3377                                 $number = substr( $number, 1 );
3378                         }
3379                         $integerPart = [];
3380                         $decimalPart = [];
3381                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3382                         preg_match( "/\d+/", $number, $integerPart );
3383                         preg_match( "/\.\d*/", $number, $decimalPart );
3384                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3385                         if ( $groupedNumber === $number ) {
3386                                 // the string does not have any number part. Eg: .12345
3387                                 return $sign . $groupedNumber;
3388                         }
3389                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3390                         while ( $start > 0 ) {
3391                                 $match = $matches[0][$numMatches - 1];
3392                                 $matchLen = strlen( $match );
3393                                 $start = $end - $matchLen;
3394                                 if ( $start < 0 ) {
3395                                         $start = 0;
3396                                 }
3397                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3398                                 $end = $start;
3399                                 if ( $numMatches > 1 ) {
3400                                         // use the last pattern for the rest of the number
3401                                         $numMatches--;
3402                                 }
3403                                 if ( $start > 0 ) {
3404                                         $groupedNumber = "," . $groupedNumber;
3405                                 }
3406                         }
3407                         return $sign . $groupedNumber;
3408                 }
3409         }
3410
3411         /**
3412          * @return string
3413          */
3414         function digitGroupingPattern() {
3415                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3416         }
3417
3418         /**
3419          * @return array
3420          */
3421         function digitTransformTable() {
3422                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3423         }
3424
3425         /**
3426          * @return array
3427          */
3428         function separatorTransformTable() {
3429                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3430         }
3431
3432         /**
3433          * @return int|null
3434          */
3435         function minimumGroupingDigits() {
3436                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3437         }
3438
3439         /**
3440          * Take a list of strings and build a locale-friendly comma-separated
3441          * list, using the local comma-separator message.
3442          * The last two strings are chained with an "and".
3443          *
3444          * @param string[] $list
3445          * @return string
3446          */
3447         public function listToText( array $list ) {
3448                 $itemCount = count( $list );
3449                 if ( $itemCount < 1 ) {
3450                         return '';
3451                 }
3452                 $text = array_pop( $list );
3453                 if ( $itemCount > 1 ) {
3454                         $and = $this->msg( 'and' )->escaped();
3455                         $space = $this->msg( 'word-separator' )->escaped();
3456                         $comma = '';
3457                         if ( $itemCount > 2 ) {
3458                                 $comma = $this->msg( 'comma-separator' )->escaped();
3459                         }
3460                         $text = implode( $comma, $list ) . $and . $space . $text;
3461                 }
3462                 return $text;
3463         }
3464
3465         /**
3466          * Take a list of strings and build a locale-friendly comma-separated
3467          * list, using the local comma-separator message.
3468          * @param string[] $list Array of strings to put in a comma list
3469          * @return string
3470          */
3471         function commaList( array $list ) {
3472                 return implode(
3473                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3474                         $list
3475                 );
3476         }
3477
3478         /**
3479          * Take a list of strings and build a locale-friendly semicolon-separated
3480          * list, using the local semicolon-separator message.
3481          * @param string[] $list Array of strings to put in a semicolon list
3482          * @return string
3483          */
3484         function semicolonList( array $list ) {
3485                 return implode(
3486                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3487                         $list
3488                 );
3489         }
3490
3491         /**
3492          * Same as commaList, but separate it with the pipe instead.
3493          * @param string[] $list Array of strings to put in a pipe list
3494          * @return string
3495          */
3496         function pipeList( array $list ) {
3497                 return implode(
3498                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3499                         $list
3500                 );
3501         }
3502
3503         /**
3504          * Truncate a string to a specified length in bytes, appending an optional
3505          * string (e.g. for ellipsis)
3506          *
3507          * If $length is negative, the string will be truncated from the beginning
3508          *
3509          * @since 1.31
3510          *
3511          * @param string $string String to truncate
3512          * @param int $length Maximum length in bytes
3513          * @param string $ellipsis String to append to the end of truncated text
3514          * @param bool $adjustLength Subtract length of ellipsis from $length
3515          *
3516          * @return string
3517          */
3518         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3519                 return $this->truncateInternal(
3520                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3521                 );
3522         }
3523
3524         /**
3525          * Truncate a string to a specified number of characters, appending an optional
3526          * string (e.g. for ellipsis).
3527          *
3528          * This provides multibyte version of truncate() method of this class, suitable for truncation
3529          * based on number of characters, instead of number of bytes.
3530          *
3531          * If $length is negative, the string will be truncated from the beginning.
3532          *
3533          * @since 1.31
3534          *
3535          * @param string $string String to truncate
3536          * @param int $length Maximum number of characters
3537          * @param string $ellipsis String to append to the end of truncated text
3538          * @param bool $adjustLength Subtract length of ellipsis from $length
3539          *
3540          * @return string
3541          */
3542         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3543                 // Passing encoding to mb_strlen and mb_substr is optional.
3544                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3545                 // explicit specification of encoding is skipped.
3546                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3547                 return $this->truncateInternal(
3548                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3549                 );
3550         }
3551
3552         /**
3553          * Internal method used for truncation. This method abstracts text truncation into
3554          * one common method, allowing users to provide length measurement function and
3555          * function for finding substring.
3556          *
3557          * For usages, see truncateForDatabase and truncateForVisual.
3558          *
3559          * @param string $string String to truncate
3560          * @param int $length Maximum length of final text
3561          * @param string $ellipsis String to append to the end of truncated text
3562          * @param bool $adjustLength Subtract length of ellipsis from $length
3563          * @param callable $measureLength Callable function used for determining the length of text
3564          * @param callable $getSubstring Callable function used for getting the substrings
3565          *
3566          * @return string
3567          */
3568         private function truncateInternal(
3569                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3570         ) {
3571                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3572                         throw new InvalidArgumentException( 'Invalid callback provided' );
3573                 }
3574
3575                 # Check if there is no need to truncate
3576                 if ( $measureLength( $string ) <= abs( $length ) ) {
3577                         return $string; // no need to truncate
3578                 }
3579
3580                 # Use the localized ellipsis character
3581                 if ( $ellipsis == '...' ) {
3582                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3583                 }
3584                 if ( $length == 0 ) {
3585                         return $ellipsis; // convention
3586                 }
3587
3588                 $stringOriginal = $string;
3589                 # If ellipsis length is >= $length then we can't apply $adjustLength
3590                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3591                         $string = $ellipsis; // this can be slightly unexpected
3592                 # Otherwise, truncate and add ellipsis...
3593                 } else {
3594                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3595                         if ( $length > 0 ) {
3596                                 $length -= $ellipsisLength;
3597                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3598                                 $string = $this->removeBadCharLast( $string );
3599                                 $string = rtrim( $string );
3600                                 $string = $string . $ellipsis;
3601                         } else {
3602                                 $length += $ellipsisLength;
3603                                 $string = $getSubstring( $string, $length ); // ...xyz
3604                                 $string = $this->removeBadCharFirst( $string );
3605                                 $string = ltrim( $string );
3606                                 $string = $ellipsis . $string;
3607                         }
3608                 }
3609
3610                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3611                 # This check is *not* redundant if $adjustLength, due to the single case where
3612                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3613                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3614                         return $string;
3615                 } else {
3616                         return $stringOriginal;
3617                 }
3618         }
3619
3620         /**
3621          * Remove bytes that represent an incomplete Unicode character
3622          * at the end of string (e.g. bytes of the char are missing)
3623          *
3624          * @param string $string
3625          * @return string
3626          */
3627         protected function removeBadCharLast( $string ) {
3628                 if ( $string != '' ) {
3629                         $char = ord( $string[strlen( $string ) - 1] );
3630                         $m = [];
3631                         if ( $char >= 0xc0 ) {
3632                                 # We got the first byte only of a multibyte char; remove it.
3633                                 $string = substr( $string, 0, -1 );
3634                         } elseif ( $char >= 0x80 &&
3635                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3636                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3637                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3638                         ) {
3639                                 # We chopped in the middle of a character; remove it
3640                                 $string = $m[1];
3641                         }
3642                 }
3643                 return $string;
3644         }
3645
3646         /**
3647          * Remove bytes that represent an incomplete Unicode character
3648          * at the start of string (e.g. bytes of the char are missing)
3649          *
3650          * @param string $string
3651          * @return string
3652          */
3653         protected function removeBadCharFirst( $string ) {
3654                 if ( $string != '' ) {
3655                         $char = ord( $string[0] );
3656                         if ( $char >= 0x80 && $char < 0xc0 ) {
3657                                 # We chopped in the middle of a character; remove the whole thing
3658                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3659                         }
3660                 }
3661                 return $string;
3662         }
3663
3664         /**
3665          * Truncate a string of valid HTML to a specified length in bytes,
3666          * appending an optional string (e.g. for ellipses), and return valid HTML
3667          *
3668          * This is only intended for styled/linked text, such as HTML with
3669          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3670          * Also, this will not detect things like "display:none" CSS.
3671          *
3672          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3673          *
3674          * @param string $text HTML string to truncate
3675          * @param int $length (zero/positive) Maximum length (including ellipses)
3676          * @param string $ellipsis String to append to the truncated text
3677          * @return string
3678          */
3679         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3680                 # Use the localized ellipsis character
3681                 if ( $ellipsis == '...' ) {
3682                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3683                 }
3684                 # Check if there is clearly no need to truncate
3685                 if ( $length <= 0 ) {
3686                         return $ellipsis; // no text shown, nothing to format (convention)
3687                 } elseif ( strlen( $text ) <= $length ) {
3688                         return $text; // string short enough even *with* HTML (short-circuit)
3689                 }
3690
3691                 $dispLen = 0; // innerHTML legth so far
3692                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3693                 $tagType = 0; // 0-open, 1-close
3694                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3695                 $entityState = 0; // 0-not entity, 1-entity
3696                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3697                 $openTags = []; // open tag stack
3698                 $maybeState = null; // possible truncation state
3699
3700                 $textLen = strlen( $text );
3701                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3702                 for ( $pos = 0; true; ++$pos ) {
3703                         # Consider truncation once the display length has reached the maximim.
3704                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3705                         # Check that we're not in the middle of a bracket/entity...
3706                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3707                                 if ( !$testingEllipsis ) {
3708                                         $testingEllipsis = true;
3709                                         # Save where we are; we will truncate here unless there turn out to
3710                                         # be so few remaining characters that truncation is not necessary.
3711                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3712                                                 $maybeState = [ $ret, $openTags ]; // save state
3713                                         }
3714                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3715                                         # String in fact does need truncation, the truncation point was OK.
3716                                         list( $ret, $openTags ) = $maybeState; // reload state
3717                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3718                                         $ret .= $ellipsis; // add ellipsis
3719                                         break;
3720                                 }
3721                         }
3722                         if ( $pos >= $textLen ) {
3723                                 break; // extra iteration just for above checks
3724                         }
3725
3726                         # Read the next char...
3727                         $ch = $text[$pos];
3728                         $lastCh = $pos ? $text[$pos - 1] : '';
3729                         $ret .= $ch; // add to result string
3730                         if ( $ch == '<' ) {
3731                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3732                                 $entityState = 0; // for bad HTML
3733                                 $bracketState = 1; // tag started (checking for backslash)
3734                         } elseif ( $ch == '>' ) {
3735                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3736                                 $entityState = 0; // for bad HTML
3737                                 $bracketState = 0; // out of brackets
3738                         } elseif ( $bracketState == 1 ) {
3739                                 if ( $ch == '/' ) {
3740                                         $tagType = 1; // close tag (e.g. "</span>")
3741                                 } else {
3742                                         $tagType = 0; // open tag (e.g. "<span>")
3743                                         $tag .= $ch;
3744                                 }
3745                                 $bracketState = 2; // building tag name
3746                         } elseif ( $bracketState == 2 ) {
3747                                 if ( $ch != ' ' ) {
3748                                         $tag .= $ch;
3749                                 } else {
3750                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3751                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3752                                 }
3753                         } elseif ( $bracketState == 0 ) {
3754                                 if ( $entityState ) {
3755                                         if ( $ch == ';' ) {
3756                                                 $entityState = 0;
3757                                                 $dispLen++; // entity is one displayed char
3758                                         }
3759                                 } else {
3760                                         if ( $neLength == 0 && !$maybeState ) {
3761                                                 // Save state without $ch. We want to *hit* the first
3762                                                 // display char (to get tags) but not *use* it if truncating.
3763                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3764                                         }
3765                                         if ( $ch == '&' ) {
3766                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3767                                         } else {
3768                                                 $dispLen++; // this char is displayed
3769                                                 // Add the next $max display text chars after this in one swoop...
3770                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3771                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3772                                                 $dispLen += $skipped;
3773                                                 $pos += $skipped;
3774                                         }
3775                                 }
3776                         }
3777                 }
3778                 // Close the last tag if left unclosed by bad HTML
3779                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3780                 while ( count( $openTags ) > 0 ) {
3781                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3782                 }
3783                 return $ret;
3784         }
3785
3786         /**
3787          * truncateHtml() helper function
3788          * like strcspn() but adds the skipped chars to $ret
3789          *
3790          * @param string $ret
3791          * @param string $text
3792          * @param string $search
3793          * @param int $start
3794          * @param null|int $len
3795          * @return int
3796          */
3797         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3798                 if ( $len === null ) {
3799                         $len = -1; // -1 means "no limit" for strcspn
3800                 } elseif ( $len < 0 ) {
3801                         $len = 0; // sanity
3802                 }
3803                 $skipCount = 0;
3804                 if ( $start < strlen( $text ) ) {
3805                         $skipCount = strcspn( $text, $search, $start, $len );
3806                         $ret .= substr( $text, $start, $skipCount );
3807                 }
3808                 return $skipCount;
3809         }
3810
3811         /**
3812          * truncateHtml() helper function
3813          * (a) push or pop $tag from $openTags as needed
3814          * (b) clear $tag value
3815          * @param string &$tag Current HTML tag name we are looking at
3816          * @param int $tagType (0-open tag, 1-close tag)
3817          * @param string $lastCh Character before the '>' that ended this tag
3818          * @param array &$openTags Open tag stack (not accounting for $tag)
3819          */
3820         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3821                 $tag = ltrim( $tag );
3822                 if ( $tag != '' ) {
3823                         if ( $tagType == 0 && $lastCh != '/' ) {
3824                                 $openTags[] = $tag; // tag opened (didn't close itself)
3825                         } elseif ( $tagType == 1 ) {
3826                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3827                                         array_pop( $openTags ); // tag closed
3828                                 }
3829                         }
3830                         $tag = '';
3831                 }
3832         }
3833
3834         /**
3835          * Grammatical transformations, needed for inflected languages
3836          * Invoked by putting {{grammar:case|word}} in a message
3837          *
3838          * @param string $word
3839          * @param string $case
3840          * @return string
3841          */
3842         function convertGrammar( $word, $case ) {
3843                 global $wgGrammarForms;
3844                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3845                         return $wgGrammarForms[$this->getCode()][$case][$word];
3846                 }
3847
3848                 $grammarTransformations = $this->getGrammarTransformations();
3849
3850                 if ( isset( $grammarTransformations[$case] ) ) {
3851                         $forms = $grammarTransformations[$case];
3852
3853                         // Some names of grammar rules are aliases for other rules.
3854                         // In such cases the value is a string rather than object,
3855                         // so load the actual rules.
3856                         if ( is_string( $forms ) ) {
3857                                 $forms = $grammarTransformations[$forms];
3858                         }
3859
3860                         foreach ( array_values( $forms ) as $rule ) {
3861                                 $form = $rule[0];
3862
3863                                 if ( $form === '@metadata' ) {
3864                                         continue;
3865                                 }
3866
3867                                 $replacement = $rule[1];
3868
3869                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3870                                 $patternMatches = preg_match( $regex, $word );
3871
3872                                 if ( $patternMatches === false ) {
3873                                         wfLogWarning(
3874                                                 'An error occurred while processing grammar. ' .
3875                                                 "Word: '$word'. Regex: /$form/."
3876                                         );
3877                                 } elseif ( $patternMatches === 1 ) {
3878                                         $word = preg_replace( $regex, $replacement, $word );
3879
3880                                         break;
3881                                 }
3882                         }
3883                 }
3884
3885                 return $word;
3886         }
3887
3888         /**
3889          * Get the grammar forms for the content language
3890          * @return array Array of grammar forms
3891          * @since 1.20
3892          */
3893         function getGrammarForms() {
3894                 global $wgGrammarForms;
3895                 if ( isset( $wgGrammarForms[$this->getCode()] )
3896                         && is_array( $wgGrammarForms[$this->getCode()] )
3897                 ) {
3898                         return $wgGrammarForms[$this->getCode()];
3899                 }
3900
3901                 return [];
3902         }
3903
3904         /**
3905          * Get the grammar transformations data for the language.
3906          * Used like grammar forms, with {{GRAMMAR}} and cases,
3907          * but uses pairs of regexes and replacements instead of code.
3908          *
3909          * @return array[] Array of grammar transformations.
3910          * @throws MWException
3911          * @since 1.28
3912          */
3913         public function getGrammarTransformations() {
3914                 $languageCode = $this->getCode();
3915
3916                 if ( self::$grammarTransformations === null ) {
3917                         self::$grammarTransformations = new MapCacheLRU( 10 );
3918                 }
3919
3920                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3921                         return self::$grammarTransformations->get( $languageCode );
3922                 }
3923
3924                 $data = [];
3925
3926                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3927                 if ( is_readable( $grammarDataFile ) ) {
3928                         $data = FormatJson::decode(
3929                                 file_get_contents( $grammarDataFile ),
3930                                 true
3931                         );
3932
3933                         if ( $data === null ) {
3934                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3935                         }
3936
3937                         self::$grammarTransformations->set( $languageCode, $data );
3938                 }
3939
3940                 return $data;
3941         }
3942
3943         /**
3944          * Provides an alternative text depending on specified gender.
3945          * Usage {{gender:username|masculine|feminine|unknown}}.
3946          * username is optional, in which case the gender of current user is used,
3947          * but only in (some) interface messages; otherwise default gender is used.
3948          *
3949          * If no forms are given, an empty string is returned. If only one form is
3950          * given, it will be returned unconditionally. These details are implied by
3951          * the caller and cannot be overridden in subclasses.
3952          *
3953          * If three forms are given, the default is to use the third (unknown) form.
3954          * If fewer than three forms are given, the default is to use the first (masculine) form.
3955          * These details can be overridden in subclasses.
3956          *
3957          * @param string $gender
3958          * @param array $forms
3959          *
3960          * @return string
3961          */
3962         function gender( $gender, $forms ) {
3963                 if ( !count( $forms ) ) {
3964                         return '';
3965                 }
3966                 $forms = $this->preConvertPlural( $forms, 2 );
3967                 if ( $gender === 'male' ) {
3968                         return $forms[0];
3969                 }
3970                 if ( $gender === 'female' ) {
3971                         return $forms[1];
3972                 }
3973                 return $forms[2] ?? $forms[0];
3974         }
3975
3976         /**
3977          * Plural form transformations, needed for some languages.
3978          * For example, there are 3 form of plural in Russian and Polish,
3979          * depending on "count mod 10". See [[w:Plural]]
3980          * For English it is pretty simple.
3981          *
3982          * Invoked by putting {{plural:count|wordform1|wordform2}}
3983          * or {{plural:count|wordform1|wordform2|wordform3}}
3984          *
3985          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3986          *
3987          * @param int $count Non-localized number
3988          * @param array $forms Different plural forms
3989          * @return string Correct form of plural for $count in this language
3990          */
3991         function convertPlural( $count, $forms ) {
3992                 // Handle explicit n=pluralform cases
3993                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3994                 if ( is_string( $forms ) ) {
3995                         return $forms;
3996                 }
3997                 if ( !count( $forms ) ) {
3998                         return '';
3999                 }
4000
4001                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
4002                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4003                 return $forms[$pluralForm];
4004         }
4005
4006         /**
4007          * Handles explicit plural forms for Language::convertPlural()
4008          *
4009          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4010          * If an explicitly defined plural form matches the $count, then
4011          * string value returned, otherwise array returned for further consideration
4012          * by CLDR rules or overridden convertPlural().
4013          *
4014          * @since 1.23
4015          *
4016          * @param int $count Non-localized number
4017          * @param array $forms Different plural forms
4018          *
4019          * @return array|string
4020          */
4021         protected function handleExplicitPluralForms( $count, array $forms ) {
4022                 foreach ( $forms as $index => $form ) {
4023                         if ( preg_match( '/\d+=/i', $form ) ) {
4024                                 $pos = strpos( $form, '=' );
4025                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4026                                         return substr( $form, $pos + 1 );
4027                                 }
4028                                 unset( $forms[$index] );
4029                         }
4030                 }
4031                 return array_values( $forms );
4032         }
4033
4034         /**
4035          * Checks that convertPlural was given an array and pads it to requested
4036          * amount of forms by copying the last one.
4037          *
4038          * @param array $forms Array of forms given to convertPlural
4039          * @param int $count How many forms should there be at least
4040          * @return array Padded array of forms or an exception if not an array
4041          */
4042         protected function preConvertPlural( /* Array */ $forms, $count ) {
4043                 while ( count( $forms ) < $count ) {
4044                         $forms[] = $forms[count( $forms ) - 1];
4045                 }
4046                 return $forms;
4047         }
4048
4049         /**
4050          * Wraps argument with unicode control characters for directionality safety
4051          *
4052          * This solves the problem where directionality-neutral characters at the edge of
4053          * the argument string get interpreted with the wrong directionality from the
4054          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4055          *
4056          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4057          * directionality of the argument string, using the BIDI algorithm's own "First
4058          * strong directional codepoint" rule. Essentially, this works round the fact that
4059          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4060          * direction inference). The latter is cleaner but still not widely supported.
4061          *
4062          * @param string $text Text to wrap
4063          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4064          */
4065         public function embedBidi( $text = '' ) {
4066                 $dir = self::strongDirFromContent( $text );
4067                 if ( $dir === 'ltr' ) {
4068                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4069                         return self::$lre . $text . self::$pdf;
4070                 }
4071                 if ( $dir === 'rtl' ) {
4072                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4073                         return self::$rle . $text . self::$pdf;
4074                 }
4075                 // No strong directionality: do not wrap
4076                 return $text;
4077         }
4078
4079         /**
4080          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4081          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4082          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4083          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4084          * match up with it.
4085          *
4086          * @param string $str The validated block duration in English
4087          * @param User|null $user User object to use timezone from or null for $wgUser
4088          * @param int $now Current timestamp, for formatting relative block durations
4089          * @return string Somehow translated block duration
4090          * @see LanguageFi.php for example implementation
4091          */
4092         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4093                 $duration = SpecialBlock::getSuggestedDurations( $this );
4094                 foreach ( $duration as $show => $value ) {
4095                         if ( strcmp( $str, $value ) == 0 ) {
4096                                 return htmlspecialchars( trim( $show ) );
4097                         }
4098                 }
4099
4100                 if ( wfIsInfinity( $str ) ) {
4101                         foreach ( $duration as $show => $value ) {
4102                                 if ( wfIsInfinity( $value ) ) {
4103                                         return htmlspecialchars( trim( $show ) );
4104                                 }
4105                         }
4106                 }
4107
4108                 // If all else fails, return a standard duration or timestamp description.
4109                 $time = strtotime( $str, $now );
4110                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4111                         return $str;
4112                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4113                         // The result differs based on current time, so the difference
4114                         // is a fixed duration length.
4115                         return $this->formatDuration( $time - $now );
4116                 } else { // It's an absolute timestamp.
4117                         if ( $time === 0 ) {
4118                                 // wfTimestamp() handles 0 as current time instead of epoch.
4119                                 $time = '19700101000000';
4120                         }
4121                         if ( $user ) {
4122                                 return $this->userTimeAndDate( $time, $user );
4123                         }
4124                         return $this->timeanddate( $time );
4125                 }
4126         }
4127
4128         /**
4129          * languages like Chinese need to be segmented in order for the diff
4130          * to be of any use
4131          *
4132          * @param string $text
4133          * @return string
4134          */
4135         public function segmentForDiff( $text ) {
4136                 return $text;
4137         }
4138
4139         /**
4140          * and unsegment to show the result
4141          *
4142          * @param string $text
4143          * @return string
4144          */
4145         public function unsegmentForDiff( $text ) {
4146                 return $text;
4147         }
4148
4149         /**
4150          * Return the LanguageConverter used in the Language
4151          *
4152          * @since 1.19
4153          * @return LanguageConverter
4154          */
4155         public function getConverter() {
4156                 return $this->mConverter;
4157         }
4158
4159         /**
4160          * convert text to a variant
4161          *
4162          * @param string $text text to convert
4163          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4164          *      variant (if logged in), or the project default variant
4165          * @return string the converted string
4166          */
4167         public function autoConvert( $text, $variant = false ) {
4168                 return $this->mConverter->autoConvert( $text, $variant );
4169         }
4170
4171         /**
4172          * convert text to all supported variants
4173          *
4174          * @param string $text
4175          * @return array
4176          */
4177         public function autoConvertToAllVariants( $text ) {
4178                 return $this->mConverter->autoConvertToAllVariants( $text );
4179         }
4180
4181         /**
4182          * convert text to different variants of a language.
4183          *
4184          * @warning Glossary state is maintained between calls. This means
4185          *  if you pass unescaped text to this method it can cause an XSS
4186          *  in later calls to this method, even if the later calls have properly
4187          *  escaped the input. Never feed this method user controlled text that
4188          *  is not properly escaped!
4189          * @param string $text Content that has been already escaped for use in HTML
4190          * @return string HTML
4191          */
4192         public function convert( $text ) {
4193                 return $this->mConverter->convert( $text );
4194         }
4195
4196         /**
4197          * Convert a Title object to a string in the preferred variant
4198          *
4199          * @param Title $title
4200          * @return string
4201          */
4202         public function convertTitle( $title ) {
4203                 return $this->mConverter->convertTitle( $title );
4204         }
4205
4206         /**
4207          * Convert a namespace index to a string in the preferred variant
4208          *
4209          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4210          * @param string|null $variant variant to convert to, or null to use the user's preferred
4211          *      variant (if logged in), or the project default variant
4212          * @return string a string representation of the namespace
4213          */
4214         public function convertNamespace( $ns, $variant = null ) {
4215                 return $this->mConverter->convertNamespace( $ns, $variant );
4216         }
4217
4218         /**
4219          * Check if this is a language with variants
4220          *
4221          * @return bool
4222          */
4223         public function hasVariants() {
4224                 return count( $this->getVariants() ) > 1;
4225         }
4226
4227         /**
4228          * Strict check if the language has the specific variant.
4229          *
4230          * Compare to LanguageConverter::validateVariant() which does a more
4231          * lenient check and attempts to coerce the given code to a valid one.
4232          *
4233          * @since 1.19
4234          * @param string $variant
4235          * @return bool
4236          */
4237         public function hasVariant( $variant ) {
4238                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4239         }
4240
4241         /**
4242          * Perform output conversion on a string, and encode for safe HTML output.
4243          * @param string $text Text to be converted
4244          * @param bool $isTitle Whether this conversion is for the article title
4245          * @return string
4246          * @todo this should get integrated somewhere sane
4247          */
4248         public function convertHtml( $text, $isTitle = false ) {
4249                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4250         }
4251
4252         /**
4253          * @param string $key
4254          * @return string
4255          */
4256         public function convertCategoryKey( $key ) {
4257                 return $this->mConverter->convertCategoryKey( $key );
4258         }
4259
4260         /**
4261          * Get the list of variants supported by this language
4262          * see sample implementation in LanguageZh.php
4263          *
4264          * @return string[] An array of language codes
4265          */
4266         public function getVariants() {
4267                 return $this->mConverter->getVariants();
4268         }
4269
4270         /**
4271          * @return string
4272          */
4273         public function getPreferredVariant() {
4274                 return $this->mConverter->getPreferredVariant();
4275         }
4276
4277         /**
4278          * @return string
4279          */
4280         public function getDefaultVariant() {
4281                 return $this->mConverter->getDefaultVariant();
4282         }
4283
4284         /**
4285          * @return string
4286          */
4287         public function getURLVariant() {
4288                 return $this->mConverter->getURLVariant();
4289         }
4290
4291         /**
4292          * If a language supports multiple variants, it is
4293          * possible that non-existing link in one variant
4294          * actually exists in another variant. this function
4295          * tries to find it. See e.g. LanguageZh.php
4296          * The input parameters may be modified upon return
4297          *
4298          * @param string &$link The name of the link
4299          * @param Title &$nt The title object of the link
4300          * @param bool $ignoreOtherCond To disable other conditions when
4301          *   we need to transclude a template or update a category's link
4302          */
4303         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4304                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4305         }
4306
4307         /**
4308          * returns language specific options used by User::getPageRenderHash()
4309          * for example, the preferred language variant
4310          *
4311          * @return string
4312          */
4313         function getExtraHashOptions() {
4314                 return $this->mConverter->getExtraHashOptions();
4315         }
4316
4317         /**
4318          * For languages that support multiple variants, the title of an
4319          * article may be displayed differently in different variants. this
4320          * function returns the apporiate title defined in the body of the article.
4321          *
4322          * @return string
4323          */
4324         public function getParsedTitle() {
4325                 return $this->mConverter->getParsedTitle();
4326         }
4327
4328         /**
4329          * Refresh the cache of conversion tables when
4330          * MediaWiki:Conversiontable* is updated.
4331          *
4332          * @param Title $title The Title of the page being updated
4333          */
4334         public function updateConversionTable( Title $title ) {
4335                 $this->mConverter->updateConversionTable( $title );
4336         }
4337
4338         /**
4339          * Prepare external link text for conversion. When the text is
4340          * a URL, it shouldn't be converted, and it'll be wrapped in
4341          * the "raw" tag (-{R| }-) to prevent conversion.
4342          *
4343          * This function is called "markNoConversion" for historical
4344          * reasons *BUT DIFFERS SIGNIFICANTLY* from
4345          * LanguageConverter::markNoConversion(), with which it is easily
4346          * confused.
4347          *
4348          * @param string $text Text to be used for external link
4349          * @param bool $noParse Wrap it without confirming it's a real URL first
4350          * @return string The tagged text
4351          * @deprecated since 1.32, use LanguageConverter::markNoConversion()
4352          *  instead.
4353          */
4354         public function markNoConversion( $text, $noParse = false ) {
4355                 wfDeprecated( __METHOD__, '1.32' );
4356                 // Excluding protocal-relative URLs may avoid many false positives.
4357                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4358                         return $this->mConverter->markNoConversion( $text );
4359                 } else {
4360                         return $text;
4361                 }
4362         }
4363
4364         /**
4365          * A regular expression to match legal word-trailing characters
4366          * which should be merged onto a link of the form [[foo]]bar.
4367          *
4368          * @return string
4369          */
4370         public function linkTrail() {
4371                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4372         }
4373
4374         /**
4375          * A regular expression character set to match legal word-prefixing
4376          * characters which should be merged onto a link of the form foo[[bar]].
4377          *
4378          * @return string
4379          */
4380         public function linkPrefixCharset() {
4381                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4382         }
4383
4384         /**
4385          * Get the "parent" language which has a converter to convert a "compatible" language
4386          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4387          *
4388          * @return Language|null
4389          * @since 1.22
4390          */
4391         public function getParentLanguage() {
4392                 if ( $this->mParentLanguage !== false ) {
4393                         return $this->mParentLanguage;
4394                 }
4395
4396                 $code = explode( '-', $this->getCode() )[0];
4397                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4398                         $this->mParentLanguage = null;
4399                         return null;
4400                 }
4401                 $lang = self::factory( $code );
4402                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4403                         $this->mParentLanguage = null;
4404                         return null;
4405                 }
4406
4407                 $this->mParentLanguage = $lang;
4408                 return $lang;
4409         }
4410
4411         /**
4412          * Compare with an other language object
4413          *
4414          * @since 1.28
4415          * @param Language $lang
4416          * @return bool
4417          */
4418         public function equals( Language $lang ) {
4419                 return $lang === $this || $lang->getCode() === $this->mCode;
4420         }
4421
4422         /**
4423          * Get the internal language code for this language object
4424          *
4425          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4426          * htmlspecialchars() or similar
4427          *
4428          * @return string
4429          */
4430         public function getCode() {
4431                 return $this->mCode;
4432         }
4433
4434         /**
4435          * Get the code in BCP 47 format which we can use
4436          * inside of html lang="" tags.
4437          *
4438          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4439          * htmlspecialchars() or similar.
4440          *
4441          * @since 1.19
4442          * @return string
4443          */
4444         public function getHtmlCode() {
4445                 if ( is_null( $this->mHtmlCode ) ) {
4446                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4447                 }
4448                 return $this->mHtmlCode;
4449         }
4450
4451         /**
4452          * @param string $code
4453          * @deprecated since 1.32, use Language::factory to create a new object instead.
4454          */
4455         public function setCode( $code ) {
4456                 $this->mCode = $code;
4457                 // Ensure we don't leave incorrect cached data lying around
4458                 $this->mHtmlCode = null;
4459                 $this->mParentLanguage = false;
4460         }
4461
4462         /**
4463          * Get the language code from a file name. Inverse of getFileName()
4464          * @param string $filename $prefix . $languageCode . $suffix
4465          * @param string $prefix Prefix before the language code
4466          * @param string $suffix Suffix after the language code
4467          * @return string Language code, or false if $prefix or $suffix isn't found
4468          */
4469         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4470                 $m = null;
4471                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4472                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4473                 if ( !count( $m ) ) {
4474                         return false;
4475                 }
4476                 return str_replace( '_', '-', strtolower( $m[1] ) );
4477         }
4478
4479         /**
4480          * @param string $code
4481          * @param bool $fallback Whether we're going through language fallback chain
4482          * @return string Name of the language class
4483          */
4484         public static function classFromCode( $code, $fallback = true ) {
4485                 if ( $fallback && $code == 'en' ) {
4486                         return 'Language';
4487                 } else {
4488                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4489                 }
4490         }
4491
4492         /**
4493          * Get the name of a file for a certain language code
4494          * @param string $prefix Prepend this to the filename
4495          * @param string $code Language code
4496          * @param string $suffix Append this to the filename
4497          * @throws MWException
4498          * @return string $prefix . $mangledCode . $suffix
4499          */
4500         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4501                 if ( !self::isValidBuiltInCode( $code ) ) {
4502                         throw new MWException( "Invalid language code \"$code\"" );
4503                 }
4504
4505                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4506         }
4507
4508         /**
4509          * @param string $code
4510          * @return string
4511          */
4512         public static function getMessagesFileName( $code ) {
4513                 global $IP;
4514                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4515                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4516                 return $file;
4517         }
4518
4519         /**
4520          * @param string $code
4521          * @return string
4522          * @throws MWException
4523          * @since 1.23
4524          */
4525         public static function getJsonMessagesFileName( $code ) {
4526                 global $IP;
4527
4528                 if ( !self::isValidBuiltInCode( $code ) ) {
4529                         throw new MWException( "Invalid language code \"$code\"" );
4530                 }
4531
4532                 return "$IP/languages/i18n/$code.json";
4533         }
4534
4535         /**
4536          * Get the first fallback for a given language.
4537          *
4538          * @param string $code
4539          *
4540          * @return bool|string
4541          */
4542         public static function getFallbackFor( $code ) {
4543                 $fallbacks = self::getFallbacksFor( $code );
4544                 if ( $fallbacks ) {
4545                         return $fallbacks[0];
4546                 }
4547                 return false;
4548         }
4549
4550         /**
4551          * Get the ordered list of fallback languages.
4552          *
4553          * @since 1.19
4554          * @param string $code Language code
4555          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4556          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4557          * @throws MWException
4558          * @return array List of language codes
4559          */
4560         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4561                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4562                         return [];
4563                 }
4564                 switch ( $mode ) {
4565                         case self::MESSAGES_FALLBACKS:
4566                                 // For unknown languages, fallbackSequence returns an empty array,
4567                                 // hardcode fallback to 'en' in that case as English messages are
4568                                 // always defined.
4569                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4570                         case self::STRICT_FALLBACKS:
4571                                 // Use this mode when you don't want to fallback to English unless
4572                                 // explicitly defined, for example when you have language-variant icons
4573                                 // and an international language-independent fallback.
4574                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4575                         default:
4576                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4577                 }
4578         }
4579
4580         /**
4581          * Get the ordered list of fallback languages, ending with the fallback
4582          * language chain for the site language.
4583          *
4584          * @since 1.22
4585          * @param string $code Language code
4586          * @return array Array( fallbacks, site fallbacks )
4587          */
4588         public static function getFallbacksIncludingSiteLanguage( $code ) {
4589                 global $wgLanguageCode;
4590
4591                 // Usually, we will only store a tiny number of fallback chains, so we
4592                 // keep them in static memory.
4593                 $cacheKey = "{$code}-{$wgLanguageCode}";
4594
4595                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4596                         $fallbacks = self::getFallbacksFor( $code );
4597
4598                         // Append the site's fallback chain, including the site language itself
4599                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4600                         array_unshift( $siteFallbacks, $wgLanguageCode );
4601
4602                         // Eliminate any languages already included in the chain
4603                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4604
4605                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4606                 }
4607                 return self::$fallbackLanguageCache[$cacheKey];
4608         }
4609
4610         /**
4611          * Get all messages for a given language
4612          * WARNING: this may take a long time. If you just need all message *keys*
4613          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4614          *
4615          * @param string $code
4616          *
4617          * @return array
4618          */
4619         public static function getMessagesFor( $code ) {
4620                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4621         }
4622
4623         /**
4624          * Get a message for a given language
4625          *
4626          * @param string $key
4627          * @param string $code
4628          *
4629          * @return string
4630          */
4631         public static function getMessageFor( $key, $code ) {
4632                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4633         }
4634
4635         /**
4636          * Get all message keys for a given language. This is a faster alternative to
4637          * array_keys( Language::getMessagesFor( $code ) )
4638          *
4639          * @since 1.19
4640          * @param string $code Language code
4641          * @return array Array of message keys (strings)
4642          */
4643         public static function getMessageKeysFor( $code ) {
4644                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4645         }
4646
4647         /**
4648          * @param string $talk
4649          * @return mixed
4650          */
4651         function fixVariableInNamespace( $talk ) {
4652                 if ( strpos( $talk, '$1' ) === false ) {
4653                         return $talk;
4654                 }
4655
4656                 global $wgMetaNamespace;
4657                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4658
4659                 # Allow grammar transformations
4660                 # Allowing full message-style parsing would make simple requests
4661                 # such as action=raw much more expensive than they need to be.
4662                 # This will hopefully cover most cases.
4663                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4664                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4665                 return str_replace( ' ', '_', $talk );
4666         }
4667
4668         /**
4669          * @param string $m
4670          * @return string
4671          */
4672         function replaceGrammarInNamespace( $m ) {
4673                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4674         }
4675
4676         /**
4677          * Decode an expiry (block, protection, etc) which has come from the DB
4678          *
4679          * @param string $expiry Database expiry String
4680          * @param bool|int $format True to process using language functions, or TS_ constant
4681          *     to return the expiry in a given timestamp
4682          * @param string $infinity If $format is not true, use this string for infinite expiry
4683          * @return string
4684          * @since 1.18
4685          */
4686         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4687                 static $dbInfinity;
4688                 if ( $dbInfinity === null ) {
4689                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4690                 }
4691
4692                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4693                         return $format === true
4694                                 ? $this->getMessageFromDB( 'infiniteblock' )
4695                                 : $infinity;
4696                 } else {
4697                         return $format === true
4698                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4699                                 : wfTimestamp( $format, $expiry );
4700                 }
4701         }
4702
4703         /**
4704          * Formats a time given in seconds into a string representation of that time.
4705          *
4706          * @param int|float $seconds
4707          * @param array $format An optional argument that formats the returned string in different ways:
4708          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4709          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4710          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4711          *     and friends.
4712          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4713          *     or 'avoidminutes'.
4714          * @return string
4715          */
4716         function formatTimePeriod( $seconds, $format = [] ) {
4717                 if ( !is_array( $format ) ) {
4718                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4719                 }
4720                 if ( !isset( $format['avoid'] ) ) {
4721                         $format['avoid'] = false;
4722                 }
4723                 if ( !isset( $format['noabbrevs'] ) ) {
4724                         $format['noabbrevs'] = false;
4725                 }
4726                 $secondsMsg = wfMessage(
4727                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4728                 $minutesMsg = wfMessage(
4729                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4730                 $hoursMsg = wfMessage(
4731                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4732                 $daysMsg = wfMessage(
4733                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4734
4735                 if ( round( $seconds * 10 ) < 100 ) {
4736                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4737                         $s = $secondsMsg->params( $s )->text();
4738                 } elseif ( round( $seconds ) < 60 ) {
4739                         $s = $this->formatNum( round( $seconds ) );
4740                         $s = $secondsMsg->params( $s )->text();
4741                 } elseif ( round( $seconds ) < 3600 ) {
4742                         $minutes = floor( $seconds / 60 );
4743                         $secondsPart = round( fmod( $seconds, 60 ) );
4744                         if ( $secondsPart == 60 ) {
4745                                 $secondsPart = 0;
4746                                 $minutes++;
4747                         }
4748                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4749                         $s .= ' ';
4750                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4751                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4752                         $hours = floor( $seconds / 3600 );
4753                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4754                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4755                         if ( $secondsPart == 60 ) {
4756                                 $secondsPart = 0;
4757                                 $minutes++;
4758                         }
4759                         if ( $minutes == 60 ) {
4760                                 $minutes = 0;
4761                                 $hours++;
4762                         }
4763                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4764                         $s .= ' ';
4765                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4766                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4767                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4768                         }
4769                 } else {
4770                         $days = floor( $seconds / 86400 );
4771                         if ( $format['avoid'] === 'avoidminutes' ) {
4772                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4773                                 if ( $hours == 24 ) {
4774                                         $hours = 0;
4775                                         $days++;
4776                                 }
4777                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4778                                 $s .= ' ';
4779                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4780                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4781                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4782                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4783                                 if ( $minutes == 60 ) {
4784                                         $minutes = 0;
4785                                         $hours++;
4786                                 }
4787                                 if ( $hours == 24 ) {
4788                                         $hours = 0;
4789                                         $days++;
4790                                 }
4791                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4792                                 $s .= ' ';
4793                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4794                                 $s .= ' ';
4795                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4796                         } else {
4797                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4798                                 $s .= ' ';
4799                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4800                         }
4801                 }
4802                 return $s;
4803         }
4804
4805         /**
4806          * Format a bitrate for output, using an appropriate
4807          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4808          *   the magnitude in question.
4809          *
4810          * This use base 1000. For base 1024 use formatSize(), for another base
4811          * see formatComputingNumbers().
4812          *
4813          * @param int $bps
4814          * @return string
4815          */
4816         function formatBitrate( $bps ) {
4817                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4818         }
4819
4820         /**
4821          * @param int $size Size of the unit
4822          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4823          * @param string $messageKey Message key to be uesd
4824          * @return string
4825          */
4826         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4827                 if ( $size <= 0 ) {
4828                         return str_replace( '$1', $this->formatNum( $size ),
4829                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4830                         );
4831                 }
4832                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4833                 $index = 0;
4834
4835                 $maxIndex = count( $sizes ) - 1;
4836                 while ( $size >= $boundary && $index < $maxIndex ) {
4837                         $index++;
4838                         $size /= $boundary;
4839                 }
4840
4841                 // For small sizes no decimal places necessary
4842                 $round = 0;
4843                 if ( $index > 1 ) {
4844                         // For MB and bigger two decimal places are smarter
4845                         $round = 2;
4846                 }
4847                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4848
4849                 $size = round( $size, $round );
4850                 $text = $this->getMessageFromDB( $msg );
4851                 return str_replace( '$1', $this->formatNum( $size ), $text );
4852         }
4853
4854         /**
4855          * Format a size in bytes for output, using an appropriate
4856          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4857          *
4858          * This method use base 1024. For base 1000 use formatBitrate(), for
4859          * another base see formatComputingNumbers()
4860          *
4861          * @param int $size Size to format
4862          * @return string Plain text (not HTML)
4863          */
4864         function formatSize( $size ) {
4865                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4866         }
4867
4868         /**
4869          * Make a list item, used by various special pages
4870          *
4871          * @param string $page Page link
4872          * @param string $details HTML safe text between brackets
4873          * @param bool $oppositedm Add the direction mark opposite to your
4874          *   language, to display text properly
4875          * @return string HTML escaped
4876          */
4877         function specialList( $page, $details, $oppositedm = true ) {
4878                 if ( !$details ) {
4879                         return $page;
4880                 }
4881
4882                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4883                 return $page .
4884                         $dirmark .
4885                         $this->msg( 'word-separator' )->escaped() .
4886                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4887         }
4888
4889         /**
4890          * Generate (prev x| next x) (20|50|100...) type links for paging
4891          *
4892          * @param Title $title Title object to link
4893          * @param int $offset
4894          * @param int $limit
4895          * @param array $query Optional URL query parameter string
4896          * @param bool $atend Optional param for specified if this is the last page
4897          * @return string
4898          */
4899         public function viewPrevNext( Title $title, $offset, $limit,
4900                 array $query = [], $atend = false
4901         ) {
4902                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4903
4904                 # Make 'previous' link
4905                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4906                 if ( $offset > 0 ) {
4907                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4908                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4909                 } else {
4910                         $plink = htmlspecialchars( $prev );
4911                 }
4912
4913                 # Make 'next' link
4914                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4915                 if ( $atend ) {
4916                         $nlink = htmlspecialchars( $next );
4917                 } else {
4918                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4919                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4920                 }
4921
4922                 # Make links to set number of items per page
4923                 $numLinks = [];
4924                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4925                         $numLinks[] = $this->numLink( $title, $offset, $num,
4926                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4927                 }
4928
4929                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4930                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4931         }
4932
4933         /**
4934          * Helper function for viewPrevNext() that generates links
4935          *
4936          * @param Title $title Title object to link
4937          * @param int $offset
4938          * @param int $limit
4939          * @param array $query Extra query parameters
4940          * @param string $link Text to use for the link; will be escaped
4941          * @param string $tooltipMsg Name of the message to use as tooltip
4942          * @param string $class Value of the "class" attribute of the link
4943          * @return string HTML fragment
4944          */
4945         private function numLink( Title $title, $offset, $limit, array $query, $link,
4946                 $tooltipMsg, $class
4947         ) {
4948                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4949                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4950                         ->numParams( $limit )->text();
4951
4952                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4953                         'title' => $tooltip, 'class' => $class ], $link );
4954         }
4955
4956         /**
4957          * Get the conversion rule title, if any.
4958          *
4959          * @return string
4960          */
4961         public function getConvRuleTitle() {
4962                 return $this->mConverter->getConvRuleTitle();
4963         }
4964
4965         /**
4966          * Get the compiled plural rules for the language
4967          * @since 1.20
4968          * @return array Associative array with plural form, and plural rule as key-value pairs
4969          */
4970         public function getCompiledPluralRules() {
4971                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4972                 $fallbacks = self::getFallbacksFor( $this->mCode );
4973                 if ( !$pluralRules ) {
4974                         foreach ( $fallbacks as $fallbackCode ) {
4975                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4976                                 if ( $pluralRules ) {
4977                                         break;
4978                                 }
4979                         }
4980                 }
4981                 return $pluralRules;
4982         }
4983
4984         /**
4985          * Get the plural rules for the language
4986          * @since 1.20
4987          * @return array Associative array with plural form number and plural rule as key-value pairs
4988          */
4989         public function getPluralRules() {
4990                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4991                 $fallbacks = self::getFallbacksFor( $this->mCode );
4992                 if ( !$pluralRules ) {
4993                         foreach ( $fallbacks as $fallbackCode ) {
4994                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4995                                 if ( $pluralRules ) {
4996                                         break;
4997                                 }
4998                         }
4999                 }
5000                 return $pluralRules;
5001         }
5002
5003         /**
5004          * Get the plural rule types for the language
5005          * @since 1.22
5006          * @return array Associative array with plural form number and plural rule type as key-value pairs
5007          */
5008         public function getPluralRuleTypes() {
5009                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
5010                 $fallbacks = self::getFallbacksFor( $this->mCode );
5011                 if ( !$pluralRuleTypes ) {
5012                         foreach ( $fallbacks as $fallbackCode ) {
5013                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
5014                                 if ( $pluralRuleTypes ) {
5015                                         break;
5016                                 }
5017                         }
5018                 }
5019                 return $pluralRuleTypes;
5020         }
5021
5022         /**
5023          * Find the index number of the plural rule appropriate for the given number
5024          * @param int $number
5025          * @return int The index number of the plural rule
5026          */
5027         public function getPluralRuleIndexNumber( $number ) {
5028                 $pluralRules = $this->getCompiledPluralRules();
5029                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5030                 return $form;
5031         }
5032
5033         /**
5034          * Find the plural rule type appropriate for the given number
5035          * For example, if the language is set to Arabic, getPluralType(5) should
5036          * return 'few'.
5037          * @since 1.22
5038          * @param int $number
5039          * @return string The name of the plural rule type, e.g. one, two, few, many
5040          */
5041         public function getPluralRuleType( $number ) {
5042                 $index = $this->getPluralRuleIndexNumber( $number );
5043                 $pluralRuleTypes = $this->getPluralRuleTypes();
5044                 return $pluralRuleTypes[$index] ?? 'other';
5045         }
5046 }