languages/Language.php

   1 <?php
   2 /**
   3  * Internationalisation code.
   4  * See https://www.mediawiki.org/wiki/Special:MyLanguage/Localisation for more information.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @ingroup Language
  23  */
  24
  25 /**
  26  * @defgroup Language Language
  27  */
  28
  29 use CLDRPluralRuleParser\Evaluator;
  30
  31 /**
  32  * Internationalisation code
  33  * @ingroup Language
  34  */
  35 class Language {
  36         /**
  37          * Return autonyms in fetchLanguageName(s).
  38          * @since 1.32
  39          */
  40         const AS_AUTONYMS = null;
  41
  42         /**
  43          * Return all known languages in fetchLanguageName(s).
  44          * @since 1.32
  45          */
  46         const ALL = 'all';
  47
  48         /**
  49          * Return in fetchLanguageName(s) only the languages for which we have at
  50          * least some localisation.
  51          * @since 1.32
  52          */
  53         const SUPPORTED = 'mwfile';
  54
  55         /**
  56          * @var LanguageConverter
  57          */
  58         public $mConverter;
  59
  60         public $mVariants, $mCode, $mLoaded = false;
  61         public $mMagicExtensions = [], $mMagicHookDone = false;
  62         private $mHtmlCode = null, $mParentLanguage = false;
  63
  64         public $dateFormatStrings = [];
  65         public $mExtendedSpecialPageAliases;
  66
  67         /** @var array|null */
  68         protected $namespaceNames;
  69         protected $mNamespaceIds, $namespaceAliases;
  70
  71         /**
  72          * ReplacementArray object caches
  73          */
  74         public $transformData = [];
  75
  76         /**
  77          * @var LocalisationCache
  78          */
  79         static public $dataCache;
  80
  81         static public $mLangObjCache = [];
  82
  83         /**
  84          * Return a fallback chain for messages in getFallbacksFor
  85          * @since 1.32
  86          */
  87         const MESSAGES_FALLBACKS = 0;
  88
  89         /**
  90          * Return a strict fallback chain in getFallbacksFor
  91          * @since 1.32
  92          */
  93         const STRICT_FALLBACKS = 1;
  94
  95         static public $mWeekdayMsgs = [
  96                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  97                 'friday', 'saturday'
  98         ];
  99
 100         static public $mWeekdayAbbrevMsgs = [
 101                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 102         ];
 103
 104         static public $mMonthMsgs = [
 105                 'january', 'february', 'march', 'april', 'may_long', 'june',
 106                 'july', 'august', 'september', 'october', 'november',
 107                 'december'
 108         ];
 109         static public $mMonthGenMsgs = [
 110                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 111                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 112                 'december-gen'
 113         ];
 114         static public $mMonthAbbrevMsgs = [
 115                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 116                 'sep', 'oct', 'nov', 'dec'
 117         ];
 118
 119         static public $mIranianCalendarMonthMsgs = [
 120                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 121                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 122                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 123                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 124         ];
 125
 126         static public $mHebrewCalendarMonthMsgs = [
 127                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 128                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 129                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 130                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 131                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 132         ];
 133
 134         static public $mHebrewCalendarMonthGenMsgs = [
 135                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 136                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 137                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 138                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 139                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 140         ];
 141
 142         static public $mHijriCalendarMonthMsgs = [
 143                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 144                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 145                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 146                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 147         ];
 148
 149         /**
 150          * @since 1.20
 151          * @var array
 152          */
 153         static public $durationIntervals = [
 154                 'millennia' => 31556952000,
 155                 'centuries' => 3155695200,
 156                 'decades' => 315569520,
 157                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 158                 'weeks' => 604800,
 159                 'days' => 86400,
 160                 'hours' => 3600,
 161                 'minutes' => 60,
 162                 'seconds' => 1,
 163         ];
 164
 165         /**
 166          * Cache for language fallbacks.
 167          * @see Language::getFallbacksIncludingSiteLanguage
 168          * @since 1.21
 169          * @var array
 170          */
 171         static private $fallbackLanguageCache = [];
 172
 173         /**
 174          * Cache for grammar rules data
 175          * @var MapCacheLRU|null
 176          */
 177         static private $grammarTransformations;
 178
 179         /**
 180          * Cache for language names
 181          * @var HashBagOStuff|null
 182          */
 183         static private $languageNameCache;
 184
 185         /**
 186          * Unicode directional formatting characters, for embedBidi()
 187          */
 188         static private $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 189         static private $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 190         static private $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 191
 192         /**
 193          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 194          * - in group 1 if it is LTR
 195          * - in group 2 if it is RTL
 196          * Does not match if there is no strong directionality codepoint.
 197          *
 198          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 199          *
 200          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 201          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 202          */
 203         // @codeCoverageIgnoreStart
 204         // phpcs:ignore Generic.Files.LineLength
 205         static private $strongDirRegex = '/(?:([\x{41}-\x{5a}\x{61}-\x{7a}\x{aa}\x{b5}\x{ba}\x{c0}-\x{d6}\x{d8}-\x{f6}\x{f8}-\x{2b8}\x{2bb}-\x{2c1}\x{2d0}\x{2d1}\x{2e0}-\x{2e4}\x{2ee}\x{370}-\x{373}\x{376}\x{377}\x{37a}-\x{37d}\x{37f}\x{386}\x{388}-\x{38a}\x{38c}\x{38e}-\x{3a1}\x{3a3}-\x{3f5}\x{3f7}-\x{482}\x{48a}-\x{52f}\x{531}-\x{556}\x{559}-\x{55f}\x{561}-\x{587}\x{589}\x{903}-\x{939}\x{93b}\x{93d}-\x{940}\x{949}-\x{94c}\x{94e}-\x{950}\x{958}-\x{961}\x{964}-\x{980}\x{982}\x{983}\x{985}-\x{98c}\x{98f}\x{990}\x{993}-\x{9a8}\x{9aa}-\x{9b0}\x{9b2}\x{9b6}-\x{9b9}\x{9bd}-\x{9c0}\x{9c7}\x{9c8}\x{9cb}\x{9cc}\x{9ce}\x{9d7}\x{9dc}\x{9dd}\x{9df}-\x{9e1}\x{9e6}-\x{9f1}\x{9f4}-\x{9fa}\x{a03}\x{a05}-\x{a0a}\x{a0f}\x{a10}\x{a13}-\x{a28}\x{a2a}-\x{a30}\x{a32}\x{a33}\x{a35}\x{a36}\x{a38}\x{a39}\x{a3e}-\x{a40}\x{a59}-\x{a5c}\x{a5e}\x{a66}-\x{a6f}\x{a72}-\x{a74}\x{a83}\x{a85}-\x{a8d}\x{a8f}-\x{a91}\x{a93}-\x{aa8}\x{aaa}-\x{ab0}\x{ab2}\x{ab3}\x{ab5}-\x{ab9}\x{abd}-\x{ac0}\x{ac9}\x{acb}\x{acc}\x{ad0}\x{ae0}\x{ae1}\x{ae6}-\x{af0}\x{af9}\x{b02}\x{b03}\x{b05}-\x{b0c}\x{b0f}\x{b10}\x{b13}-\x{b28}\x{b2a}-\x{b30}\x{b32}\x{b33}\x{b35}-\x{b39}\x{b3d}\x{b3e}\x{b40}\x{b47}\x{b48}\x{b4b}\x{b4c}\x{b57}\x{b5c}\x{b5d}\x{b5f}-\x{b61}\x{b66}-\x{b77}\x{b83}\x{b85}-\x{b8a}\x{b8e}-\x{b90}\x{b92}-\x{b95}\x{b99}\x{b9a}\x{b9c}\x{b9e}\x{b9f}\x{ba3}\x{ba4}\x{ba8}-\x{baa}\x{bae}-\x{bb9}\x{bbe}\x{bbf}\x{bc1}\x{bc2}\x{bc6}-\x{bc8}\x{bca}-\x{bcc}\x{bd0}\x{bd7}\x{be6}-\x{bf2}\x{c01}-\x{c03}\x{c05}-\x{c0c}\x{c0e}-\x{c10}\x{c12}-\x{c28}\x{c2a}-\x{c39}\x{c3d}\x{c41}-\x{c44}\x{c58}-\x{c5a}\x{c60}\x{c61}\x{c66}-\x{c6f}\x{c7f}\x{c82}\x{c83}\x{c85}-\x{c8c}\x{c8e}-\x{c90}\x{c92}-\x{ca8}\x{caa}-\x{cb3}\x{cb5}-\x{cb9}\x{cbd}-\x{cc4}\x{cc6}-\x{cc8}\x{cca}\x{ccb}\x{cd5}\x{cd6}\x{cde}\x{ce0}\x{ce1}\x{ce6}-\x{cef}\x{cf1}\x{cf2}\x{d02}\x{d03}\x{d05}-\x{d0c}\x{d0e}-\x{d10}\x{d12}-\x{d3a}\x{d3d}-\x{d40}\x{d46}-\x{d48}\x{d4a}-\x{d4c}\x{d4e}\x{d57}\x{d5f}-\x{d61}\x{d66}-\x{d75}\x{d79}-\x{d7f}\x{d82}\x{d83}\x{d85}-\x{d96}\x{d9a}-\x{db1}\x{db3}-\x{dbb}\x{dbd}\x{dc0}-\x{dc6}\x{dcf}-\x{dd1}\x{dd8}-\x{ddf}\x{de6}-\x{def}\x{df2}-\x{df4}\x{e01}-\x{e30}\x{e32}\x{e33}\x{e40}-\x{e46}\x{e4f}-\x{e5b}\x{e81}\x{e82}\x{e84}\x{e87}\x{e88}\x{e8a}\x{e8d}\x{e94}-\x{e97}\x{e99}-\x{e9f}\x{ea1}-\x{ea3}\x{ea5}\x{ea7}\x{eaa}\x{eab}\x{ead}-\x{eb0}\x{eb2}\x{eb3}\x{ebd}\x{ec0}-\x{ec4}\x{ec6}\x{ed0}-\x{ed9}\x{edc}-\x{edf}\x{f00}-\x{f17}\x{f1a}-\x{f34}\x{f36}\x{f38}\x{f3e}-\x{f47}\x{f49}-\x{f6c}\x{f7f}\x{f85}\x{f88}-\x{f8c}\x{fbe}-\x{fc5}\x{fc7}-\x{fcc}\x{fce}-\x{fda}\x{1000}-\x{102c}\x{1031}\x{1038}\x{103b}\x{103c}\x{103f}-\x{1057}\x{105a}-\x{105d}\x{1061}-\x{1070}\x{1075}-\x{1081}\x{1083}\x{1084}\x{1087}-\x{108c}\x{108e}-\x{109c}\x{109e}-\x{10c5}\x{10c7}\x{10cd}\x{10d0}-\x{1248}\x{124a}-\x{124d}\x{1250}-\x{1256}\x{1258}\x{125a}-\x{125d}\x{1260}-\x{1288}\x{128a}-\x{128d}\x{1290}-\x{12b0}\x{12b2}-\x{12b5}\x{12b8}-\x{12be}\x{12c0}\x{12c2}-\x{12c5}\x{12c8}-\x{12d6}\x{12d8}-\x{1310}\x{1312}-\x{1315}\x{1318}-\x{135a}\x{1360}-\x{137c}\x{1380}-\x{138f}\x{13a0}-\x{13f5}\x{13f8}-\x{13fd}\x{1401}-\x{167f}\x{1681}-\x{169a}\x{16a0}-\x{16f8}\x{1700}-\x{170c}\x{170e}-\x{1711}\x{1720}-\x{1731}\x{1735}\x{1736}\x{1740}-\x{1751}\x{1760}-\x{176c}\x{176e}-\x{1770}\x{1780}-\x{17b3}\x{17b6}\x{17be}-\x{17c5}\x{17c7}\x{17c8}\x{17d4}-\x{17da}\x{17dc}\x{17e0}-\x{17e9}\x{1810}-\x{1819}\x{1820}-\x{1877}\x{1880}-\x{18a8}\x{18aa}\x{18b0}-\x{18f5}\x{1900}-\x{191e}\x{1923}-\x{1926}\x{1929}-\x{192b}\x{1930}\x{1931}\x{1933}-\x{1938}\x{1946}-\x{196d}\x{1970}-\x{1974}\x{1980}-\x{19ab}\x{19b0}-\x{19c9}\x{19d0}-\x{19da}\x{1a00}-\x{1a16}\x{1a19}\x{1a1a}\x{1a1e}-\x{1a55}\x{1a57}\x{1a61}\x{1a63}\x{1a64}\x{1a6d}-\x{1a72}\x{1a80}-\x{1a89}\x{1a90}-\x{1a99}\x{1aa0}-\x{1aad}\x{1b04}-\x{1b33}\x{1b35}\x{1b3b}\x{1b3d}-\x{1b41}\x{1b43}-\x{1b4b}\x{1b50}-\x{1b6a}\x{1b74}-\x{1b7c}\x{1b82}-\x{1ba1}\x{1ba6}\x{1ba7}\x{1baa}\x{1bae}-\x{1be5}\x{1be7}\x{1bea}-\x{1bec}\x{1bee}\x{1bf2}\x{1bf3}\x{1bfc}-\x{1c2b}\x{1c34}\x{1c35}\x{1c3b}-\x{1c49}\x{1c4d}-\x{1c7f}\x{1cc0}-\x{1cc7}\x{1cd3}\x{1ce1}\x{1ce9}-\x{1cec}\x{1cee}-\x{1cf3}\x{1cf5}\x{1cf6}\x{1d00}-\x{1dbf}\x{1e00}-\x{1f15}\x{1f18}-\x{1f1d}\x{1f20}-\x{1f45}\x{1f48}-\x{1f4d}\x{1f50}-\x{1f57}\x{1f59}\x{1f5b}\x{1f5d}\x{1f5f}-\x{1f7d}\x{1f80}-\x{1fb4}\x{1fb6}-\x{1fbc}\x{1fbe}\x{1fc2}-\x{1fc4}\x{1fc6}-\x{1fcc}\x{1fd0}-\x{1fd3}\x{1fd6}-\x{1fdb}\x{1fe0}-\x{1fec}\x{1ff2}-\x{1ff4}\x{1ff6}-\x{1ffc}\x{200e}\x{2071}\x{207f}\x{2090}-\x{209c}\x{2102}\x{2107}\x{210a}-\x{2113}\x{2115}\x{2119}-\x{211d}\x{2124}\x{2126}\x{2128}\x{212a}-\x{212d}\x{212f}-\x{2139}\x{213c}-\x{213f}\x{2145}-\x{2149}\x{214e}\x{214f}\x{2160}-\x{2188}\x{2336}-\x{237a}\x{2395}\x{249c}-\x{24e9}\x{26ac}\x{2800}-\x{28ff}\x{2c00}-\x{2c2e}\x{2c30}-\x{2c5e}\x{2c60}-\x{2ce4}\x{2ceb}-\x{2cee}\x{2cf2}\x{2cf3}\x{2d00}-\x{2d25}\x{2d27}\x{2d2d}\x{2d30}-\x{2d67}\x{2d6f}\x{2d70}\x{2d80}-\x{2d96}\x{2da0}-\x{2da6}\x{2da8}-\x{2dae}\x{2db0}-\x{2db6}\x{2db8}-\x{2dbe}\x{2dc0}-\x{2dc6}\x{2dc8}-\x{2dce}\x{2dd0}-\x{2dd6}\x{2dd8}-\x{2dde}\x{3005}-\x{3007}\x{3021}-\x{3029}\x{302e}\x{302f}\x{3031}-\x{3035}\x{3038}-\x{303c}\x{3041}-\x{3096}\x{309d}-\x{309f}\x{30a1}-\x{30fa}\x{30fc}-\x{30ff}\x{3105}-\x{312d}\x{3131}-\x{318e}\x{3190}-\x{31ba}\x{31f0}-\x{321c}\x{3220}-\x{324f}\x{3260}-\x{327b}\x{327f}-\x{32b0}\x{32c0}-\x{32cb}\x{32d0}-\x{32fe}\x{3300}-\x{3376}\x{337b}-\x{33dd}\x{33e0}-\x{33fe}\x{3400}-\x{4db5}\x{4e00}-\x{9fd5}\x{a000}-\x{a48c}\x{a4d0}-\x{a60c}\x{a610}-\x{a62b}\x{a640}-\x{a66e}\x{a680}-\x{a69d}\x{a6a0}-\x{a6ef}\x{a6f2}-\x{a6f7}\x{a722}-\x{a787}\x{a789}-\x{a7ad}\x{a7b0}-\x{a7b7}\x{a7f7}-\x{a801}\x{a803}-\x{a805}\x{a807}-\x{a80a}\x{a80c}-\x{a824}\x{a827}\x{a830}-\x{a837}\x{a840}-\x{a873}\x{a880}-\x{a8c3}\x{a8ce}-\x{a8d9}\x{a8f2}-\x{a8fd}\x{a900}-\x{a925}\x{a92e}-\x{a946}\x{a952}\x{a953}\x{a95f}-\x{a97c}\x{a983}-\x{a9b2}\x{a9b4}\x{a9b5}\x{a9ba}\x{a9bb}\x{a9bd}-\x{a9cd}\x{a9cf}-\x{a9d9}\x{a9de}-\x{a9e4}\x{a9e6}-\x{a9fe}\x{aa00}-\x{aa28}\x{aa2f}\x{aa30}\x{aa33}\x{aa34}\x{aa40}-\x{aa42}\x{aa44}-\x{aa4b}\x{aa4d}\x{aa50}-\x{aa59}\x{aa5c}-\x{aa7b}\x{aa7d}-\x{aaaf}\x{aab1}\x{aab5}\x{aab6}\x{aab9}-\x{aabd}\x{aac0}\x{aac2}\x{aadb}-\x{aaeb}\x{aaee}-\x{aaf5}\x{ab01}-\x{ab06}\x{ab09}-\x{ab0e}\x{ab11}-\x{ab16}\x{ab20}-\x{ab26}\x{ab28}-\x{ab2e}\x{ab30}-\x{ab65}\x{ab70}-\x{abe4}\x{abe6}\x{abe7}\x{abe9}-\x{abec}\x{abf0}-\x{abf9}\x{ac00}-\x{d7a3}\x{d7b0}-\x{d7c6}\x{d7cb}-\x{d7fb}\x{e000}-\x{fa6d}\x{fa70}-\x{fad9}\x{fb00}-\x{fb06}\x{fb13}-\x{fb17}\x{ff21}-\x{ff3a}\x{ff41}-\x{ff5a}\x{ff66}-\x{ffbe}\x{ffc2}-\x{ffc7}\x{ffca}-\x{ffcf}\x{ffd2}-\x{ffd7}\x{ffda}-\x{ffdc}\x{10000}-\x{1000b}\x{1000d}-\x{10026}\x{10028}-\x{1003a}\x{1003c}\x{1003d}\x{1003f}-\x{1004d}\x{10050}-\x{1005d}\x{10080}-\x{100fa}\x{10100}\x{10102}\x{10107}-\x{10133}\x{10137}-\x{1013f}\x{101d0}-\x{101fc}\x{10280}-\x{1029c}\x{102a0}-\x{102d0}\x{10300}-\x{10323}\x{10330}-\x{1034a}\x{10350}-\x{10375}\x{10380}-\x{1039d}\x{1039f}-\x{103c3}\x{103c8}-\x{103d5}\x{10400}-\x{1049d}\x{104a0}-\x{104a9}\x{10500}-\x{10527}\x{10530}-\x{10563}\x{1056f}\x{10600}-\x{10736}\x{10740}-\x{10755}\x{10760}-\x{10767}\x{11000}\x{11002}-\x{11037}\x{11047}-\x{1104d}\x{11066}-\x{1106f}\x{11082}-\x{110b2}\x{110b7}\x{110b8}\x{110bb}-\x{110c1}\x{110d0}-\x{110e8}\x{110f0}-\x{110f9}\x{11103}-\x{11126}\x{1112c}\x{11136}-\x{11143}\x{11150}-\x{11172}\x{11174}-\x{11176}\x{11182}-\x{111b5}\x{111bf}-\x{111c9}\x{111cd}\x{111d0}-\x{111df}\x{111e1}-\x{111f4}\x{11200}-\x{11211}\x{11213}-\x{1122e}\x{11232}\x{11233}\x{11235}\x{11238}-\x{1123d}\x{11280}-\x{11286}\x{11288}\x{1128a}-\x{1128d}\x{1128f}-\x{1129d}\x{1129f}-\x{112a9}\x{112b0}-\x{112de}\x{112e0}-\x{112e2}\x{112f0}-\x{112f9}\x{11302}\x{11303}\x{11305}-\x{1130c}\x{1130f}\x{11310}\x{11313}-\x{11328}\x{1132a}-\x{11330}\x{11332}\x{11333}\x{11335}-\x{11339}\x{1133d}-\x{1133f}\x{11341}-\x{11344}\x{11347}\x{11348}\x{1134b}-\x{1134d}\x{11350}\x{11357}\x{1135d}-\x{11363}\x{11480}-\x{114b2}\x{114b9}\x{114bb}-\x{114be}\x{114c1}\x{114c4}-\x{114c7}\x{114d0}-\x{114d9}\x{11580}-\x{115b1}\x{115b8}-\x{115bb}\x{115be}\x{115c1}-\x{115db}\x{11600}-\x{11632}\x{1163b}\x{1163c}\x{1163e}\x{11641}-\x{11644}\x{11650}-\x{11659}\x{11680}-\x{116aa}\x{116ac}\x{116ae}\x{116af}\x{116b6}\x{116c0}-\x{116c9}\x{11700}-\x{11719}\x{11720}\x{11721}\x{11726}\x{11730}-\x{1173f}\x{118a0}-\x{118f2}\x{118ff}\x{11ac0}-\x{11af8}\x{12000}-\x{12399}\x{12400}-\x{1246e}\x{12470}-\x{12474}\x{12480}-\x{12543}\x{13000}-\x{1342e}\x{14400}-\x{14646}\x{16800}-\x{16a38}\x{16a40}-\x{16a5e}\x{16a60}-\x{16a69}\x{16a6e}\x{16a6f}\x{16ad0}-\x{16aed}\x{16af5}\x{16b00}-\x{16b2f}\x{16b37}-\x{16b45}\x{16b50}-\x{16b59}\x{16b5b}-\x{16b61}\x{16b63}-\x{16b77}\x{16b7d}-\x{16b8f}\x{16f00}-\x{16f44}\x{16f50}-\x{16f7e}\x{16f93}-\x{16f9f}\x{1b000}\x{1b001}\x{1bc00}-\x{1bc6a}\x{1bc70}-\x{1bc7c}\x{1bc80}-\x{1bc88}\x{1bc90}-\x{1bc99}\x{1bc9c}\x{1bc9f}\x{1d000}-\x{1d0f5}\x{1d100}-\x{1d126}\x{1d129}-\x{1d166}\x{1d16a}-\x{1d172}\x{1d183}\x{1d184}\x{1d18c}-\x{1d1a9}\x{1d1ae}-\x{1d1e8}\x{1d360}-\x{1d371}\x{1d400}-\x{1d454}\x{1d456}-\x{1d49c}\x{1d49e}\x{1d49f}\x{1d4a2}\x{1d4a5}\x{1d4a6}\x{1d4a9}-\x{1d4ac}\x{1d4ae}-\x{1d4b9}\x{1d4bb}\x{1d4bd}-\x{1d4c3}\x{1d4c5}-\x{1d505}\x{1d507}-\x{1d50a}\x{1d50d}-\x{1d514}\x{1d516}-\x{1d51c}\x{1d51e}-\x{1d539}\x{1d53b}-\x{1d53e}\x{1d540}-\x{1d544}\x{1d546}\x{1d54a}-\x{1d550}\x{1d552}-\x{1d6a5}\x{1d6a8}-\x{1d6da}\x{1d6dc}-\x{1d714}\x{1d716}-\x{1d74e}\x{1d750}-\x{1d788}\x{1d78a}-\x{1d7c2}\x{1d7c4}-\x{1d7cb}\x{1d800}-\x{1d9ff}\x{1da37}-\x{1da3a}\x{1da6d}-\x{1da74}\x{1da76}-\x{1da83}\x{1da85}-\x{1da8b}\x{1f110}-\x{1f12e}\x{1f130}-\x{1f169}\x{1f170}-\x{1f19a}\x{1f1e6}-\x{1f202}\x{1f210}-\x{1f23a}\x{1f240}-\x{1f248}\x{1f250}\x{1f251}\x{20000}-\x{2a6d6}\x{2a700}-\x{2b734}\x{2b740}-\x{2b81d}\x{2b820}-\x{2cea1}\x{2f800}-\x{2fa1d}\x{f0000}-\x{ffffd}\x{100000}-\x{10fffd}])|([\x{590}\x{5be}\x{5c0}\x{5c3}\x{5c6}\x{5c8}-\x{5ff}\x{7c0}-\x{7ea}\x{7f4}\x{7f5}\x{7fa}-\x{815}\x{81a}\x{824}\x{828}\x{82e}-\x{858}\x{85c}-\x{89f}\x{200f}\x{fb1d}\x{fb1f}-\x{fb28}\x{fb2a}-\x{fb4f}\x{10800}-\x{1091e}\x{10920}-\x{10a00}\x{10a04}\x{10a07}-\x{10a0b}\x{10a10}-\x{10a37}\x{10a3b}-\x{10a3e}\x{10a40}-\x{10ae4}\x{10ae7}-\x{10b38}\x{10b40}-\x{10e5f}\x{10e7f}-\x{10fff}\x{1e800}-\x{1e8cf}\x{1e8d7}-\x{1edff}\x{1ef00}-\x{1efff}\x{608}\x{60b}\x{60d}\x{61b}-\x{64a}\x{66d}-\x{66f}\x{671}-\x{6d5}\x{6e5}\x{6e6}\x{6ee}\x{6ef}\x{6fa}-\x{710}\x{712}-\x{72f}\x{74b}-\x{7a5}\x{7b1}-\x{7bf}\x{8a0}-\x{8e2}\x{fb50}-\x{fd3d}\x{fd40}-\x{fdcf}\x{fdf0}-\x{fdfc}\x{fdfe}\x{fdff}\x{fe70}-\x{fefe}\x{1ee00}-\x{1eeef}\x{1eef2}-\x{1eeff}]))/u';
 206         // @codeCoverageIgnoreEnd
 207
 208         /**
 209          * Get a cached or new language object for a given language code
 210          * @param string $code
 211          * @throws MWException
 212          * @return Language
 213          */
 214         static function factory( $code ) {
 215                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 216
 217                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 218                         $code = $wgDummyLanguageCodes[$code];
 219                 }
 220
 221                 // get the language object to process
 222                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 223
 224                 // merge the language object in to get it up front in the cache
 225                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 226                 // get rid of the oldest ones in case we have an overflow
 227                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 228
 229                 return $langObj;
 230         }
 231
 232         /**
 233          * Create a language object for a given language code
 234          * @param string $code
 235          * @param bool $fallback Whether we're going through language fallback chain
 236          * @throws MWException
 237          * @return Language
 238          */
 239         protected static function newFromCode( $code, $fallback = false ) {
 240                 if ( !self::isValidCode( $code ) ) {
 241                         throw new MWException( "Invalid language code \"$code\"" );
 242                 }
 243
 244                 if ( !self::isValidBuiltInCode( $code ) ) {
 245                         // It's not possible to customise this code with class files, so
 246                         // just return a Language object. This is to support uselang= hacks.
 247                         $lang = new Language;
 248                         $lang->setCode( $code );
 249                         return $lang;
 250                 }
 251
 252                 // Check if there is a language class for the code
 253                 $class = self::classFromCode( $code, $fallback );
 254                 // LanguageCode does not inherit Language
 255                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 256                         $lang = new $class;
 257                         return $lang;
 258                 }
 259
 260                 // Keep trying the fallback list until we find an existing class
 261                 $fallbacks = self::getFallbacksFor( $code );
 262                 foreach ( $fallbacks as $fallbackCode ) {
 263                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 264                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 265                         }
 266
 267                         $class = self::classFromCode( $fallbackCode );
 268                         if ( class_exists( $class ) ) {
 269                                 $lang = new $class;
 270                                 $lang->setCode( $code );
 271                                 return $lang;
 272                         }
 273                 }
 274
 275                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 276         }
 277
 278         /**
 279          * Intended for tests that may change configuration in a way that invalidates caches.
 280          *
 281          * @since 1.32
 282          */
 283         public static function clearCaches() {
 284                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 285                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 286                 }
 287                 self::$dataCache = null;
 288                 // Reinitialize $dataCache, since it's expected to always be available
 289                 self::getLocalisationCache();
 290                 self::$mLangObjCache = [];
 291                 self::$fallbackLanguageCache = [];
 292                 self::$grammarTransformations = null;
 293                 self::$languageNameCache = null;
 294         }
 295
 296         /**
 297          * Checks whether any localisation is available for that language tag
 298          * in MediaWiki (MessagesXx.php exists).
 299          *
 300          * @param string $code Language tag (in lower case)
 301          * @return bool Whether language is supported
 302          * @since 1.21
 303          */
 304         public static function isSupportedLanguage( $code ) {
 305                 if ( !self::isValidBuiltInCode( $code ) ) {
 306                         return false;
 307                 }
 308
 309                 if ( $code === 'qqq' ) {
 310                         return false;
 311                 }
 312
 313                 return is_readable( self::getMessagesFileName( $code ) ) ||
 314                         is_readable( self::getJsonMessagesFileName( $code ) );
 315         }
 316
 317         /**
 318          * Returns true if a language code string is a well-formed language tag
 319          * according to RFC 5646.
 320          * This function only checks well-formedness; it doesn't check that
 321          * language, script or variant codes actually exist in the repositories.
 322          *
 323          * Based on regexes by Mark Davis of the Unicode Consortium:
 324          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 325          *
 326          * @param string $code
 327          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 328          *
 329          * @return bool
 330          * @since 1.21
 331          */
 332         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 333                 $alpha = '[a-z]';
 334                 $digit = '[0-9]';
 335                 $alphanum = '[a-z0-9]';
 336                 $x = 'x'; # private use singleton
 337                 $singleton = '[a-wy-z]'; # other singleton
 338                 $s = $lenient ? '[-_]' : '-';
 339
 340                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 341                 $script = "$alpha{4}"; # ISO 15924
 342                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 343                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 344                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 345                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 346
 347                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 348                 # Since these are limited, this is safe even later changes to the registry --
 349                 # the only oddity is that it might change the type of the tag, and thus
 350                 # the results from the capturing groups.
 351                 # https://www.iana.org/assignments/language-subtag-registry
 352
 353                 $grandfathered = "en{$s}GB{$s}oed"
 354                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 355                         . "|no{$s}(?:bok|nyn)"
 356                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 357                         . "|zh{$s}min{$s}nan";
 358
 359                 $variantList = "$variant(?:$s$variant)*";
 360                 $extensionList = "$extension(?:$s$extension)*";
 361
 362                 $langtag = "(?:($language)"
 363                         . "(?:$s$script)?"
 364                         . "(?:$s$region)?"
 365                         . "(?:$s$variantList)?"
 366                         . "(?:$s$extensionList)?"
 367                         . "(?:$s$privateUse)?)";
 368
 369                 # The final breakdown, with capturing groups for each of these components
 370                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 371
 372                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 373
 374                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 375         }
 376
 377         /**
 378          * Returns true if a language code string is of a valid form, whether or
 379          * not it exists. This includes codes which are used solely for
 380          * customisation via the MediaWiki namespace.
 381          *
 382          * @param string $code
 383          *
 384          * @return bool
 385          */
 386         public static function isValidCode( $code ) {
 387                 static $cache = [];
 388                 if ( !isset( $cache[$code] ) ) {
 389                         // People think language codes are html safe, so enforce it.
 390                         // Ideally we should only allow a-zA-Z0-9-
 391                         // but, .+ and other chars are often used for {{int:}} hacks
 392                         // see bugs T39564, T39587, T38938
 393                         $cache[$code] =
 394                                 // Protect against path traversal
 395                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 396                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 397                 }
 398                 return $cache[$code];
 399         }
 400
 401         /**
 402          * Returns true if a language code is of a valid form for the purposes of
 403          * internal customisation of MediaWiki, via Messages*.php or *.json.
 404          *
 405          * @param string $code
 406          *
 407          * @throws MWException
 408          * @since 1.18
 409          * @return bool
 410          */
 411         public static function isValidBuiltInCode( $code ) {
 412                 if ( !is_string( $code ) ) {
 413                         if ( is_object( $code ) ) {
 414                                 $addmsg = " of class " . get_class( $code );
 415                         } else {
 416                                 $addmsg = '';
 417                         }
 418                         $type = gettype( $code );
 419                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 420                 }
 421
 422                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 423         }
 424
 425         /**
 426          * Returns true if a language code is an IETF tag known to MediaWiki.
 427          *
 428          * @param string $tag
 429          *
 430          * @since 1.21
 431          * @return bool
 432          */
 433         public static function isKnownLanguageTag( $tag ) {
 434                 // Quick escape for invalid input to avoid exceptions down the line
 435                 // when code tries to process tags which are not valid at all.
 436                 if ( !self::isValidBuiltInCode( $tag ) ) {
 437                         return false;
 438                 }
 439
 440                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 441                         || self::fetchLanguageName( $tag, $tag ) !== ''
 442                 ) {
 443                         return true;
 444                 }
 445
 446                 return false;
 447         }
 448
 449         /**
 450          * Get the LocalisationCache instance
 451          *
 452          * @return LocalisationCache
 453          */
 454         public static function getLocalisationCache() {
 455                 if ( is_null( self::$dataCache ) ) {
 456                         global $wgLocalisationCacheConf;
 457                         $class = $wgLocalisationCacheConf['class'];
 458                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 459                 }
 460                 return self::$dataCache;
 461         }
 462
 463         function __construct() {
 464                 $this->mConverter = new FakeConverter( $this );
 465                 // Set the code to the name of the descendant
 466                 if ( static::class === 'Language' ) {
 467                         $this->mCode = 'en';
 468                 } else {
 469                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 470                 }
 471                 self::getLocalisationCache();
 472         }
 473
 474         /**
 475          * Reduce memory usage
 476          */
 477         function __destruct() {
 478                 foreach ( $this as $name => $value ) {
 479                         unset( $this->$name );
 480                 }
 481         }
 482
 483         /**
 484          * Hook which will be called if this is the content language.
 485          * Descendants can use this to register hook functions or modify globals
 486          */
 487         function initContLang() {
 488         }
 489
 490         /**
 491          * @return array
 492          * @since 1.19
 493          */
 494         public function getFallbackLanguages() {
 495                 return self::getFallbacksFor( $this->mCode );
 496         }
 497
 498         /**
 499          * Exports $wgBookstoreListEn
 500          * @return array
 501          */
 502         public function getBookstoreList() {
 503                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 504         }
 505
 506         /**
 507          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 508          * available in localised form, it will be included in English.
 509          *
 510          * @return array
 511          */
 512         public function getNamespaces() {
 513                 if ( is_null( $this->namespaceNames ) ) {
 514                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 515
 516                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 517
 518                         $this->namespaceNames = $wgExtraNamespaces +
 519                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 520                         $this->namespaceNames += $validNamespaces;
 521
 522                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 523                         if ( $wgMetaNamespaceTalk ) {
 524                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 525                         } else {
 526                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 527                                 $this->namespaceNames[NS_PROJECT_TALK] =
 528                                         $this->fixVariableInNamespace( $talk );
 529                         }
 530
 531                         # Sometimes a language will be localised but not actually exist on this wiki.
 532                         foreach ( $this->namespaceNames as $key => $text ) {
 533                                 if ( !isset( $validNamespaces[$key] ) ) {
 534                                         unset( $this->namespaceNames[$key] );
 535                                 }
 536                         }
 537
 538                         # The above mixing may leave namespaces out of canonical order.
 539                         # Re-order by namespace ID number...
 540                         ksort( $this->namespaceNames );
 541
 542                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 543                 }
 544
 545                 return $this->namespaceNames;
 546         }
 547
 548         /**
 549          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 550          * @param array $namespaces Array of namespaces (id => name)
 551          */
 552         public function setNamespaces( array $namespaces ) {
 553                 $this->namespaceNames = $namespaces;
 554                 $this->mNamespaceIds = null;
 555         }
 556
 557         /**
 558          * Resets all of the namespace caches. Mainly used for testing
 559          */
 560         public function resetNamespaces() {
 561                 $this->namespaceNames = null;
 562                 $this->mNamespaceIds = null;
 563                 $this->namespaceAliases = null;
 564         }
 565
 566         /**
 567          * A convenience function that returns getNamespaces() with spaces instead of underscores
 568          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 569          *
 570          * @return array
 571          */
 572         public function getFormattedNamespaces() {
 573                 $ns = $this->getNamespaces();
 574                 foreach ( $ns as $k => $v ) {
 575                         $ns[$k] = strtr( $v, '_', ' ' );
 576                 }
 577                 return $ns;
 578         }
 579
 580         /**
 581          * Get a namespace value by key
 582          *
 583          * <code>
 584          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 585          * echo $mw_ns; // prints 'MediaWiki'
 586          * </code>
 587          *
 588          * @param int $index The array key of the namespace to return
 589          * @return string|bool String if the namespace value exists, otherwise false
 590          */
 591         public function getNsText( $index ) {
 592                 $ns = $this->getNamespaces();
 593                 return $ns[$index] ?? false;
 594         }
 595
 596         /**
 597          * A convenience function that returns the same thing as
 598          * getNsText() except with '_' changed to ' ', useful for
 599          * producing output.
 600          *
 601          * <code>
 602          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 603          * echo $mw_ns; // prints 'MediaWiki talk'
 604          * </code>
 605          *
 606          * @param int $index The array key of the namespace to return
 607          * @return string Namespace name without underscores (empty string if namespace does not exist)
 608          */
 609         public function getFormattedNsText( $index ) {
 610                 $ns = $this->getNsText( $index );
 611                 return strtr( $ns, '_', ' ' );
 612         }
 613
 614         /**
 615          * Returns gender-dependent namespace alias if available.
 616          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 617          * @param int $index Namespace index
 618          * @param string $gender Gender key (male, female... )
 619          * @return string
 620          * @since 1.18
 621          */
 622         public function getGenderNsText( $index, $gender ) {
 623                 global $wgExtraGenderNamespaces;
 624
 625                 $ns = $wgExtraGenderNamespaces +
 626                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 627
 628                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 629         }
 630
 631         /**
 632          * Whether this language uses gender-dependent namespace aliases.
 633          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 634          * @return bool
 635          * @since 1.18
 636          */
 637         public function needsGenderDistinction() {
 638                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 639                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 640                         // $wgExtraGenderNamespaces overrides everything
 641                         return true;
 642                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 643                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 644                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 645                         return false;
 646                 } else {
 647                         // Check what is in i18n files
 648                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 649                         return count( $aliases ) > 0;
 650                 }
 651         }
 652
 653         /**
 654          * Get a namespace key by value, case insensitive.
 655          * Only matches namespace names for the current language, not the
 656          * canonical ones defined in Namespace.php.
 657          *
 658          * @param string $text
 659          * @return int|bool An integer if $text is a valid value otherwise false
 660          */
 661         function getLocalNsIndex( $text ) {
 662                 $lctext = $this->lc( $text );
 663                 $ids = $this->getNamespaceIds();
 664                 return $ids[$lctext] ?? false;
 665         }
 666
 667         /**
 668          * @return array
 669          */
 670         public function getNamespaceAliases() {
 671                 if ( is_null( $this->namespaceAliases ) ) {
 672                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 673                         if ( !$aliases ) {
 674                                 $aliases = [];
 675                         } else {
 676                                 foreach ( $aliases as $name => $index ) {
 677                                         if ( $index === NS_PROJECT_TALK ) {
 678                                                 unset( $aliases[$name] );
 679                                                 $name = $this->fixVariableInNamespace( $name );
 680                                                 $aliases[$name] = $index;
 681                                         }
 682                                 }
 683                         }
 684
 685                         global $wgExtraGenderNamespaces;
 686                         $genders = $wgExtraGenderNamespaces +
 687                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 688                         foreach ( $genders as $index => $forms ) {
 689                                 foreach ( $forms as $alias ) {
 690                                         $aliases[$alias] = $index;
 691                                 }
 692                         }
 693
 694                         # Also add converted namespace names as aliases, to avoid confusion.
 695                         $convertedNames = [];
 696                         foreach ( $this->getVariants() as $variant ) {
 697                                 if ( $variant === $this->mCode ) {
 698                                         continue;
 699                                 }
 700                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 701                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 702                                 }
 703                         }
 704
 705                         $this->namespaceAliases = $aliases + $convertedNames;
 706
 707                         # 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          * @param string $image
 819          * @return array|null
 820          */
 821         function getImageFile( $image ) {
 822                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 823         }
 824
 825         /**
 826          * @return array
 827          * @since 1.24
 828          */
 829         public function getImageFiles() {
 830                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 831         }
 832
 833         /**
 834          * @return array
 835          */
 836         public function getExtraUserToggles() {
 837                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 838         }
 839
 840         /**
 841          * @param string $tog
 842          * @return string
 843          */
 844         function getUserToggle( $tog ) {
 845                 return $this->getMessageFromDB( "tog-$tog" );
 846         }
 847
 848         /**
 849          * Get an array of language names, indexed by code.
 850          * @param null|string $inLanguage Code of language in which to return the names
 851          *              Use self::AS_AUTONYMS for autonyms (native names)
 852          * @param string $include One of:
 853          *              self::ALL all available languages
 854          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 855          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 856          * @return array Language code => language name (sorted by key)
 857          * @since 1.20
 858          */
 859         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 860                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 861                 $cacheKey .= ":$include";
 862                 if ( self::$languageNameCache === null ) {
 863                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 864                 }
 865
 866                 $ret = self::$languageNameCache->get( $cacheKey );
 867                 if ( !$ret ) {
 868                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 869                         self::$languageNameCache->set( $cacheKey, $ret );
 870                 }
 871                 return $ret;
 872         }
 873
 874         /**
 875          * Uncached helper for fetchLanguageNames
 876          * @param null|string $inLanguage Code of language in which to return the names
 877          *              Use self::AS_AUTONYMS for autonyms (native names)
 878          * @param string $include One of:
 879          *              self::ALL all available languages
 880          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 881          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 882          * @return array Language code => language name (sorted by key)
 883          */
 884         private static function fetchLanguageNamesUncached(
 885                 $inLanguage = self::AS_AUTONYMS,
 886                 $include = 'mw'
 887         ) {
 888                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 889
 890                 // If passed an invalid language code to use, fallback to en
 891                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 892                         $inLanguage = 'en';
 893                 }
 894
 895                 $names = [];
 896
 897                 if ( $inLanguage ) {
 898                         # TODO: also include when $inLanguage is null, when this code is more efficient
 899                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 900                 }
 901
 902                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 903                 if ( $wgUsePigLatinVariant ) {
 904                         // Pig Latin (for variant development)
 905                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 906                 }
 907
 908                 foreach ( $mwNames as $mwCode => $mwName ) {
 909                         # - Prefer own MediaWiki native name when not using the hook
 910                         # - For other names just add if not added through the hook
 911                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 912                                 $names[$mwCode] = $mwName;
 913                         }
 914                 }
 915
 916                 if ( $include === self::ALL ) {
 917                         ksort( $names );
 918                         return $names;
 919                 }
 920
 921                 $returnMw = [];
 922                 $coreCodes = array_keys( $mwNames );
 923                 foreach ( $coreCodes as $coreCode ) {
 924                         $returnMw[$coreCode] = $names[$coreCode];
 925                 }
 926
 927                 if ( $include === self::SUPPORTED ) {
 928                         $namesMwFile = [];
 929                         # We do this using a foreach over the codes instead of a directory
 930                         # loop so that messages files in extensions will work correctly.
 931                         foreach ( $returnMw as $code => $value ) {
 932                                 if ( is_readable( self::getMessagesFileName( $code ) )
 933                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 934                                 ) {
 935                                         $namesMwFile[$code] = $names[$code];
 936                                 }
 937                         }
 938
 939                         ksort( $namesMwFile );
 940                         return $namesMwFile;
 941                 }
 942
 943                 ksort( $returnMw );
 944                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 945                 return $returnMw;
 946         }
 947
 948         /**
 949          * @param string $code The code of the language for which to get the name
 950          * @param null|string $inLanguage Code of language in which to return the name
 951          *   (SELF::AS_AUTONYMS for autonyms)
 952          * @param string $include See fetchLanguageNames()
 953          * @return string Language name or empty
 954          * @since 1.20
 955          */
 956         public static function fetchLanguageName(
 957                 $code,
 958                 $inLanguage = self::AS_AUTONYMS,
 959                 $include = self::ALL
 960         ) {
 961                 $code = strtolower( $code );
 962                 $array = self::fetchLanguageNames( $inLanguage, $include );
 963                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 964         }
 965
 966         /**
 967          * Get a message from the MediaWiki namespace.
 968          *
 969          * @param string $msg Message name
 970          * @return string
 971          */
 972         public function getMessageFromDB( $msg ) {
 973                 return $this->msg( $msg )->text();
 974         }
 975
 976         /**
 977          * Get message object in this language. Only for use inside this class.
 978          *
 979          * @param string $msg Message name
 980          * @return Message
 981          */
 982         protected function msg( $msg ) {
 983                 return wfMessage( $msg )->inLanguage( $this );
 984         }
 985
 986         /**
 987          * @param string $key
 988          * @return string
 989          */
 990         public function getMonthName( $key ) {
 991                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 992         }
 993
 994         /**
 995          * @return array
 996          */
 997         public function getMonthNamesArray() {
 998                 $monthNames = [ '' ];
 999                 for ( $i = 1; $i < 13; $i++ ) {
1000                         $monthNames[] = $this->getMonthName( $i );
1001                 }
1002                 return $monthNames;
1003         }
1004
1005         /**
1006          * @param string $key
1007          * @return string
1008          */
1009         public function getMonthNameGen( $key ) {
1010                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
1011         }
1012
1013         /**
1014          * @param string $key
1015          * @return string
1016          */
1017         public function getMonthAbbreviation( $key ) {
1018                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
1019         }
1020
1021         /**
1022          * @return array
1023          */
1024         public function getMonthAbbreviationsArray() {
1025                 $monthNames = [ '' ];
1026                 for ( $i = 1; $i < 13; $i++ ) {
1027                         $monthNames[] = $this->getMonthAbbreviation( $i );
1028                 }
1029                 return $monthNames;
1030         }
1031
1032         /**
1033          * @param string $key
1034          * @return string
1035          */
1036         public function getWeekdayName( $key ) {
1037                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1038         }
1039
1040         /**
1041          * @param string $key
1042          * @return string
1043          */
1044         function getWeekdayAbbreviation( $key ) {
1045                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1046         }
1047
1048         /**
1049          * @param string $key
1050          * @return string
1051          */
1052         function getIranianCalendarMonthName( $key ) {
1053                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1054         }
1055
1056         /**
1057          * @param string $key
1058          * @return string
1059          */
1060         function getHebrewCalendarMonthName( $key ) {
1061                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1062         }
1063
1064         /**
1065          * @param string $key
1066          * @return string
1067          */
1068         function getHebrewCalendarMonthNameGen( $key ) {
1069                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1070         }
1071
1072         /**
1073          * @param string $key
1074          * @return string
1075          */
1076         function getHijriCalendarMonthName( $key ) {
1077                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1078         }
1079
1080         /**
1081          * Pass through result from $dateTimeObj->format()
1082          * @param DateTime|bool|null &$dateTimeObj
1083          * @param string $ts
1084          * @param DateTimeZone|bool|null $zone
1085          * @param string $code
1086          * @return string
1087          */
1088         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1089                 if ( !$dateTimeObj ) {
1090                         $dateTimeObj = DateTime::createFromFormat(
1091                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1092                         );
1093                 }
1094                 return $dateTimeObj->format( $code );
1095         }
1096
1097         /**
1098          * This is a workalike of PHP's date() function, but with better
1099          * internationalisation, a reduced set of format characters, and a better
1100          * escaping format.
1101          *
1102          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1103          * the PHP manual for definitions. There are a number of extensions, which
1104          * start with "x":
1105          *
1106          *    xn   Do not translate digits of the next numeric format character
1107          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1108          *    xr   Use roman numerals for the next numeric format character
1109          *    xh   Use hebrew numerals for the next numeric format character
1110          *    xx   Literal x
1111          *    xg   Genitive month name
1112          *
1113          *    xij  j (day number) in Iranian calendar
1114          *    xiF  F (month name) in Iranian calendar
1115          *    xin  n (month number) in Iranian calendar
1116          *    xiy  y (two digit year) in Iranian calendar
1117          *    xiY  Y (full year) in Iranian calendar
1118          *    xit  t (days in month) in Iranian calendar
1119          *    xiz  z (day of the year) in Iranian calendar
1120          *
1121          *    xjj  j (day number) in Hebrew calendar
1122          *    xjF  F (month name) in Hebrew calendar
1123          *    xjt  t (days in month) in Hebrew calendar
1124          *    xjx  xg (genitive month name) in Hebrew calendar
1125          *    xjn  n (month number) in Hebrew calendar
1126          *    xjY  Y (full year) in Hebrew calendar
1127          *
1128          *    xmj  j (day number) in Hijri calendar
1129          *    xmF  F (month name) in Hijri calendar
1130          *    xmn  n (month number) in Hijri calendar
1131          *    xmY  Y (full year) in Hijri calendar
1132          *
1133          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1134          *                       identical to the Gregorian calendar
1135          *    xoY  Y (full year) in Minguo calendar or Juche year.
1136          *                       Months and days are identical to the
1137          *                       Gregorian calendar
1138          *    xtY  Y (full year) in Japanese nengo. Months and days are
1139          *                       identical to the Gregorian calendar
1140          *
1141          * Characters enclosed in double quotes will be considered literal (with
1142          * the quotes themselves removed). Unmatched quotes will be considered
1143          * literal quotes. Example:
1144          *
1145          * "The month is" F       => The month is January
1146          * i's"                   => 20'11"
1147          *
1148          * Backslash escaping is also supported.
1149          *
1150          * Input timestamp is assumed to be pre-normalized to the desired local
1151          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1152          * $ts is UTC if $zone is not given.
1153          *
1154          * @param string $format
1155          * @param string $ts 14-character timestamp
1156          *      YYYYMMDDHHMMSS
1157          *      01234567890123
1158          * @param DateTimeZone|null $zone Timezone of $ts
1159          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1160          * Only makes sense if $ts is the current time.
1161          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1162          *
1163          * @throws MWException
1164          * @return string
1165          */
1166         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1167                 $s = '';
1168                 $raw = false;
1169                 $roman = false;
1170                 $hebrewNum = false;
1171                 $dateTimeObj = false;
1172                 $rawToggle = false;
1173                 $iranian = false;
1174                 $hebrew = false;
1175                 $hijri = false;
1176                 $thai = false;
1177                 $minguo = false;
1178                 $tenno = false;
1179
1180                 $usedSecond = false;
1181                 $usedMinute = false;
1182                 $usedHour = false;
1183                 $usedAMPM = false;
1184                 $usedDay = false;
1185                 $usedWeek = false;
1186                 $usedMonth = false;
1187                 $usedYear = false;
1188                 $usedISOYear = false;
1189                 $usedIsLeapYear = false;
1190
1191                 $usedHebrewMonth = false;
1192                 $usedIranianMonth = false;
1193                 $usedHijriMonth = false;
1194                 $usedHebrewYear = false;
1195                 $usedIranianYear = false;
1196                 $usedHijriYear = false;
1197                 $usedTennoYear = false;
1198
1199                 if ( strlen( $ts ) !== 14 ) {
1200                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1201                 }
1202
1203                 if ( !ctype_digit( $ts ) ) {
1204                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1205                 }
1206
1207                 $formatLength = strlen( $format );
1208                 for ( $p = 0; $p < $formatLength; $p++ ) {
1209                         $num = false;
1210                         $code = $format[$p];
1211                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1212                                 $code .= $format[++$p];
1213                         }
1214
1215                         if ( ( $code === 'xi'
1216                                         || $code === 'xj'
1217                                         || $code === 'xk'
1218                                         || $code === 'xm'
1219                                         || $code === 'xo'
1220                                         || $code === 'xt' )
1221                                 && $p < $formatLength - 1 ) {
1222                                 $code .= $format[++$p];
1223                         }
1224
1225                         switch ( $code ) {
1226                                 case 'xx':
1227                                         $s .= 'x';
1228                                         break;
1229                                 case 'xn':
1230                                         $raw = true;
1231                                         break;
1232                                 case 'xN':
1233                                         $rawToggle = !$rawToggle;
1234                                         break;
1235                                 case 'xr':
1236                                         $roman = true;
1237                                         break;
1238                                 case 'xh':
1239                                         $hebrewNum = true;
1240                                         break;
1241                                 case 'xg':
1242                                         $usedMonth = true;
1243                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1244                                         break;
1245                                 case 'xjx':
1246                                         $usedHebrewMonth = true;
1247                                         if ( !$hebrew ) {
1248                                                 $hebrew = self::tsToHebrew( $ts );
1249                                         }
1250                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1251                                         break;
1252                                 case 'd':
1253                                         $usedDay = true;
1254                                         $num = substr( $ts, 6, 2 );
1255                                         break;
1256                                 case 'D':
1257                                         $usedDay = true;
1258                                         $s .= $this->getWeekdayAbbreviation(
1259                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1260                                         );
1261                                         break;
1262                                 case 'j':
1263                                         $usedDay = true;
1264                                         $num = intval( substr( $ts, 6, 2 ) );
1265                                         break;
1266                                 case 'xij':
1267                                         $usedDay = true;
1268                                         if ( !$iranian ) {
1269                                                 $iranian = self::tsToIranian( $ts );
1270                                         }
1271                                         $num = $iranian[2];
1272                                         break;
1273                                 case 'xmj':
1274                                         $usedDay = true;
1275                                         if ( !$hijri ) {
1276                                                 $hijri = self::tsToHijri( $ts );
1277                                         }
1278                                         $num = $hijri[2];
1279                                         break;
1280                                 case 'xjj':
1281                                         $usedDay = true;
1282                                         if ( !$hebrew ) {
1283                                                 $hebrew = self::tsToHebrew( $ts );
1284                                         }
1285                                         $num = $hebrew[2];
1286                                         break;
1287                                 case 'l':
1288                                         $usedDay = true;
1289                                         $s .= $this->getWeekdayName(
1290                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1291                                         );
1292                                         break;
1293                                 case 'F':
1294                                         $usedMonth = true;
1295                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1296                                         break;
1297                                 case 'xiF':
1298                                         $usedIranianMonth = true;
1299                                         if ( !$iranian ) {
1300                                                 $iranian = self::tsToIranian( $ts );
1301                                         }
1302                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1303                                         break;
1304                                 case 'xmF':
1305                                         $usedHijriMonth = true;
1306                                         if ( !$hijri ) {
1307                                                 $hijri = self::tsToHijri( $ts );
1308                                         }
1309                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1310                                         break;
1311                                 case 'xjF':
1312                                         $usedHebrewMonth = true;
1313                                         if ( !$hebrew ) {
1314                                                 $hebrew = self::tsToHebrew( $ts );
1315                                         }
1316                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1317                                         break;
1318                                 case 'm':
1319                                         $usedMonth = true;
1320                                         $num = substr( $ts, 4, 2 );
1321                                         break;
1322                                 case 'M':
1323                                         $usedMonth = true;
1324                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1325                                         break;
1326                                 case 'n':
1327                                         $usedMonth = true;
1328                                         $num = intval( substr( $ts, 4, 2 ) );
1329                                         break;
1330                                 case 'xin':
1331                                         $usedIranianMonth = true;
1332                                         if ( !$iranian ) {
1333                                                 $iranian = self::tsToIranian( $ts );
1334                                         }
1335                                         $num = $iranian[1];
1336                                         break;
1337                                 case 'xmn':
1338                                         $usedHijriMonth = true;
1339                                         if ( !$hijri ) {
1340                                                 $hijri = self::tsToHijri( $ts );
1341                                         }
1342                                         $num = $hijri[1];
1343                                         break;
1344                                 case 'xjn':
1345                                         $usedHebrewMonth = true;
1346                                         if ( !$hebrew ) {
1347                                                 $hebrew = self::tsToHebrew( $ts );
1348                                         }
1349                                         $num = $hebrew[1];
1350                                         break;
1351                                 case 'xjt':
1352                                         $usedHebrewMonth = true;
1353                                         if ( !$hebrew ) {
1354                                                 $hebrew = self::tsToHebrew( $ts );
1355                                         }
1356                                         $num = $hebrew[3];
1357                                         break;
1358                                 case 'Y':
1359                                         $usedYear = true;
1360                                         $num = substr( $ts, 0, 4 );
1361                                         break;
1362                                 case 'xiY':
1363                                         $usedIranianYear = true;
1364                                         if ( !$iranian ) {
1365                                                 $iranian = self::tsToIranian( $ts );
1366                                         }
1367                                         $num = $iranian[0];
1368                                         break;
1369                                 case 'xmY':
1370                                         $usedHijriYear = true;
1371                                         if ( !$hijri ) {
1372                                                 $hijri = self::tsToHijri( $ts );
1373                                         }
1374                                         $num = $hijri[0];
1375                                         break;
1376                                 case 'xjY':
1377                                         $usedHebrewYear = true;
1378                                         if ( !$hebrew ) {
1379                                                 $hebrew = self::tsToHebrew( $ts );
1380                                         }
1381                                         $num = $hebrew[0];
1382                                         break;
1383                                 case 'xkY':
1384                                         $usedYear = true;
1385                                         if ( !$thai ) {
1386                                                 $thai = self::tsToYear( $ts, 'thai' );
1387                                         }
1388                                         $num = $thai[0];
1389                                         break;
1390                                 case 'xoY':
1391                                         $usedYear = true;
1392                                         if ( !$minguo ) {
1393                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1394                                         }
1395                                         $num = $minguo[0];
1396                                         break;
1397                                 case 'xtY':
1398                                         $usedTennoYear = true;
1399                                         if ( !$tenno ) {
1400                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1401                                         }
1402                                         $num = $tenno[0];
1403                                         break;
1404                                 case 'y':
1405                                         $usedYear = true;
1406                                         $num = substr( $ts, 2, 2 );
1407                                         break;
1408                                 case 'xiy':
1409                                         $usedIranianYear = true;
1410                                         if ( !$iranian ) {
1411                                                 $iranian = self::tsToIranian( $ts );
1412                                         }
1413                                         $num = substr( $iranian[0], -2 );
1414                                         break;
1415                                 case 'xit':
1416                                         $usedIranianYear = true;
1417                                         if ( !$iranian ) {
1418                                                 $iranian = self::tsToIranian( $ts );
1419                                         }
1420                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1421                                         break;
1422                                 case 'xiz':
1423                                         $usedIranianYear = true;
1424                                         if ( !$iranian ) {
1425                                                 $iranian = self::tsToIranian( $ts );
1426                                         }
1427                                         $num = $iranian[3];
1428                                         break;
1429                                 case 'a':
1430                                         $usedAMPM = true;
1431                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1432                                         break;
1433                                 case 'A':
1434                                         $usedAMPM = true;
1435                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1436                                         break;
1437                                 case 'g':
1438                                         $usedHour = true;
1439                                         $h = substr( $ts, 8, 2 );
1440                                         $num = $h % 12 ? $h % 12 : 12;
1441                                         break;
1442                                 case 'G':
1443                                         $usedHour = true;
1444                                         $num = intval( substr( $ts, 8, 2 ) );
1445                                         break;
1446                                 case 'h':
1447                                         $usedHour = true;
1448                                         $h = substr( $ts, 8, 2 );
1449                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1450                                         break;
1451                                 case 'H':
1452                                         $usedHour = true;
1453                                         $num = substr( $ts, 8, 2 );
1454                                         break;
1455                                 case 'i':
1456                                         $usedMinute = true;
1457                                         $num = substr( $ts, 10, 2 );
1458                                         break;
1459                                 case 's':
1460                                         $usedSecond = true;
1461                                         $num = substr( $ts, 12, 2 );
1462                                         break;
1463                                 case 'c':
1464                                 case 'r':
1465                                         $usedSecond = true;
1466                                         // fall through
1467                                 case 'e':
1468                                 case 'O':
1469                                 case 'P':
1470                                 case 'T':
1471                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1472                                         break;
1473                                 case 'w':
1474                                 case 'N':
1475                                 case 'z':
1476                                         $usedDay = true;
1477                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1478                                         break;
1479                                 case 'W':
1480                                         $usedWeek = true;
1481                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1482                                         break;
1483                                 case 't':
1484                                         $usedMonth = true;
1485                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1486                                         break;
1487                                 case 'L':
1488                                         $usedIsLeapYear = true;
1489                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1490                                         break;
1491                                 case 'o':
1492                                         $usedISOYear = true;
1493                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1494                                         break;
1495                                 case 'U':
1496                                         $usedSecond = true;
1497                                         // fall through
1498                                 case 'I':
1499                                 case 'Z':
1500                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1501                                         break;
1502                                 case '\\':
1503                                         # Backslash escaping
1504                                         if ( $p < $formatLength - 1 ) {
1505                                                 $s .= $format[++$p];
1506                                         } else {
1507                                                 $s .= '\\';
1508                                         }
1509                                         break;
1510                                 case '"':
1511                                         # Quoted literal
1512                                         if ( $p < $formatLength - 1 ) {
1513                                                 $endQuote = strpos( $format, '"', $p + 1 );
1514                                                 if ( $endQuote === false ) {
1515                                                         # No terminating quote, assume literal "
1516                                                         $s .= '"';
1517                                                 } else {
1518                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1519                                                         $p = $endQuote;
1520                                                 }
1521                                         } else {
1522                                                 # Quote at end of string, assume literal "
1523                                                 $s .= '"';
1524                                         }
1525                                         break;
1526                                 default:
1527                                         $s .= $format[$p];
1528                         }
1529                         if ( $num !== false ) {
1530                                 if ( $rawToggle || $raw ) {
1531                                         $s .= $num;
1532                                         $raw = false;
1533                                 } elseif ( $roman ) {
1534                                         $s .= self::romanNumeral( $num );
1535                                         $roman = false;
1536                                 } elseif ( $hebrewNum ) {
1537                                         $s .= self::hebrewNumeral( $num );
1538                                         $hebrewNum = false;
1539                                 } else {
1540                                         $s .= $this->formatNum( $num, true );
1541                                 }
1542                         }
1543                 }
1544
1545                 if ( $ttl === 'unused' ) {
1546                         // No need to calculate the TTL, the caller wont use it anyway.
1547                 } elseif ( $usedSecond ) {
1548                         $ttl = 1;
1549                 } elseif ( $usedMinute ) {
1550                         $ttl = 60 - substr( $ts, 12, 2 );
1551                 } elseif ( $usedHour ) {
1552                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1553                 } elseif ( $usedAMPM ) {
1554                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1555                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1556                 } elseif (
1557                         $usedDay ||
1558                         $usedHebrewMonth ||
1559                         $usedIranianMonth ||
1560                         $usedHijriMonth ||
1561                         $usedHebrewYear ||
1562                         $usedIranianYear ||
1563                         $usedHijriYear ||
1564                         $usedTennoYear
1565                 ) {
1566                         // @todo Someone who understands the non-Gregorian calendars
1567                         // should write proper logic for them so that they don't need purged every day.
1568                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1569                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1570                 } else {
1571                         $possibleTtls = [];
1572                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1573                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1574                         if ( $usedWeek ) {
1575                                 $possibleTtls[] =
1576                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1577                                         $timeRemainingInDay;
1578                         } elseif ( $usedISOYear ) {
1579                                 // December 28th falls on the last ISO week of the year, every year.
1580                                 // The last ISO week of a year can be 52 or 53.
1581                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1582                                         'Ymd',
1583                                         substr( $ts, 0, 4 ) . '1228',
1584                                         $zone ?: new DateTimeZone( 'UTC' )
1585                                 )->format( 'W' );
1586                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1587                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1588                                 $timeRemainingInWeek =
1589                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1590                                         + $timeRemainingInDay;
1591                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1592                         }
1593
1594                         if ( $usedMonth ) {
1595                                 $possibleTtls[] =
1596                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1597                                                 substr( $ts, 6, 2 ) ) * 86400
1598                                         + $timeRemainingInDay;
1599                         } elseif ( $usedYear ) {
1600                                 $possibleTtls[] =
1601                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1602                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1603                                         + $timeRemainingInDay;
1604                         } elseif ( $usedIsLeapYear ) {
1605                                 $year = substr( $ts, 0, 4 );
1606                                 $timeRemainingInYear =
1607                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1608                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1609                                         + $timeRemainingInDay;
1610                                 $mod = $year % 4;
1611                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1612                                         // this isn't a leap year. see when the next one starts
1613                                         $nextCandidate = $year - $mod + 4;
1614                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1615                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1616                                                         $timeRemainingInYear;
1617                                         } else {
1618                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1619                                                         $timeRemainingInYear;
1620                                         }
1621                                 } else {
1622                                         // this is a leap year, so the next year isn't
1623                                         $possibleTtls[] = $timeRemainingInYear;
1624                                 }
1625                         }
1626
1627                         if ( $possibleTtls ) {
1628                                 $ttl = min( $possibleTtls );
1629                         }
1630                 }
1631
1632                 return $s;
1633         }
1634
1635         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1636         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1637
1638         /**
1639          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1640          * Gregorian dates to Iranian dates. Originally written in C, it
1641          * is released under the terms of GNU Lesser General Public
1642          * License. Conversion to PHP was performed by Niklas Laxström.
1643          *
1644          * Link: http://www.farsiweb.info/jalali/jalali.c
1645          *
1646          * @param string $ts
1647          *
1648          * @return int[]
1649          */
1650         private static function tsToIranian( $ts ) {
1651                 $gy = substr( $ts, 0, 4 ) - 1600;
1652                 $gm = substr( $ts, 4, 2 ) - 1;
1653                 $gd = substr( $ts, 6, 2 ) - 1;
1654
1655                 # Days passed from the beginning (including leap years)
1656                 $gDayNo = 365 * $gy
1657                         + floor( ( $gy + 3 ) / 4 )
1658                         - floor( ( $gy + 99 ) / 100 )
1659                         + floor( ( $gy + 399 ) / 400 );
1660
1661                 // Add days of the past months of this year
1662                 for ( $i = 0; $i < $gm; $i++ ) {
1663                         $gDayNo += self::$GREG_DAYS[$i];
1664                 }
1665
1666                 // Leap years
1667                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1668                         $gDayNo++;
1669                 }
1670
1671                 // Days passed in current month
1672                 $gDayNo += (int)$gd;
1673
1674                 $jDayNo = $gDayNo - 79;
1675
1676                 $jNp = floor( $jDayNo / 12053 );
1677                 $jDayNo %= 12053;
1678
1679                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1680                 $jDayNo %= 1461;
1681
1682                 if ( $jDayNo >= 366 ) {
1683                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1684                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1685                 }
1686
1687                 $jz = $jDayNo;
1688
1689                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1690                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1691                 }
1692
1693                 $jm = $i + 1;
1694                 $jd = $jDayNo + 1;
1695
1696                 return [ $jy, $jm, $jd, $jz ];
1697         }
1698
1699         /**
1700          * Converting Gregorian dates to Hijri dates.
1701          *
1702          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1703          *
1704          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1705          *
1706          * @param string $ts
1707          *
1708          * @return int[]
1709          */
1710         private static function tsToHijri( $ts ) {
1711                 $year = substr( $ts, 0, 4 );
1712                 $month = substr( $ts, 4, 2 );
1713                 $day = substr( $ts, 6, 2 );
1714
1715                 $zyr = $year;
1716                 $zd = $day;
1717                 $zm = $month;
1718                 $zy = $zyr;
1719
1720                 if (
1721                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1722                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1723                 ) {
1724                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1725                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1726                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1727                                         $zd - 32075;
1728                 } else {
1729                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1730                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1731                 }
1732
1733                 $zl = $zjd - 1948440 + 10632;
1734                 $zn = (int)( ( $zl - 1 ) / 10631 );
1735                 $zl = $zl - 10631 * $zn + 354;
1736                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1737                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1738                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1739                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1740                 $zm = (int)( ( 24 * $zl ) / 709 );
1741                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1742                 $zy = 30 * $zn + $zj - 30;
1743
1744                 return [ $zy, $zm, $zd ];
1745         }
1746
1747         /**
1748          * Converting Gregorian dates to Hebrew dates.
1749          *
1750          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1751          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1752          * to translate the relevant functions into PHP and release them under
1753          * GNU GPL.
1754          *
1755          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1756          * and Adar II is 14. In a non-leap year, Adar is 6.
1757          *
1758          * @param string $ts
1759          *
1760          * @return int[]
1761          */
1762         private static function tsToHebrew( $ts ) {
1763                 # Parse date
1764                 $year = substr( $ts, 0, 4 );
1765                 $month = substr( $ts, 4, 2 );
1766                 $day = substr( $ts, 6, 2 );
1767
1768                 # Calculate Hebrew year
1769                 $hebrewYear = $year + 3760;
1770
1771                 # Month number when September = 1, August = 12
1772                 $month += 4;
1773                 if ( $month > 12 ) {
1774                         # Next year
1775                         $month -= 12;
1776                         $year++;
1777                         $hebrewYear++;
1778                 }
1779
1780                 # Calculate day of year from 1 September
1781                 $dayOfYear = $day;
1782                 for ( $i = 1; $i < $month; $i++ ) {
1783                         if ( $i == 6 ) {
1784                                 # February
1785                                 $dayOfYear += 28;
1786                                 # Check if the year is leap
1787                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1788                                         $dayOfYear++;
1789                                 }
1790                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1791                                 $dayOfYear += 30;
1792                         } else {
1793                                 $dayOfYear += 31;
1794                         }
1795                 }
1796
1797                 # Calculate the start of the Hebrew year
1798                 $start = self::hebrewYearStart( $hebrewYear );
1799
1800                 # Calculate next year's start
1801                 if ( $dayOfYear <= $start ) {
1802                         # Day is before the start of the year - it is the previous year
1803                         # Next year's start
1804                         $nextStart = $start;
1805                         # Previous year
1806                         $year--;
1807                         $hebrewYear--;
1808                         # Add days since previous year's 1 September
1809                         $dayOfYear += 365;
1810                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1811                                 # Leap year
1812                                 $dayOfYear++;
1813                         }
1814                         # Start of the new (previous) year
1815                         $start = self::hebrewYearStart( $hebrewYear );
1816                 } else {
1817                         # Next year's start
1818                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1819                 }
1820
1821                 # Calculate Hebrew day of year
1822                 $hebrewDayOfYear = $dayOfYear - $start;
1823
1824                 # Difference between year's days
1825                 $diff = $nextStart - $start;
1826                 # Add 12 (or 13 for leap years) days to ignore the difference between
1827                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1828                 # difference is only about the year type
1829                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1830                         $diff += 13;
1831                 } else {
1832                         $diff += 12;
1833                 }
1834
1835                 # Check the year pattern, and is leap year
1836                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1837                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1838                 # and non-leap years
1839                 $yearPattern = $diff % 30;
1840                 # Check if leap year
1841                 $isLeap = $diff >= 30;
1842
1843                 # Calculate day in the month from number of day in the Hebrew year
1844                 # Don't check Adar - if the day is not in Adar, we will stop before;
1845                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1846                 $hebrewDay = $hebrewDayOfYear;
1847                 $hebrewMonth = 1;
1848                 $days = 0;
1849                 while ( $hebrewMonth <= 12 ) {
1850                         # Calculate days in this month
1851                         if ( $isLeap && $hebrewMonth == 6 ) {
1852                                 # Adar in a leap year
1853                                 if ( $isLeap ) {
1854                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1855                                         $days = 30;
1856                                         if ( $hebrewDay <= $days ) {
1857                                                 # Day in Adar I
1858                                                 $hebrewMonth = 13;
1859                                         } else {
1860                                                 # Subtract the days of Adar I
1861                                                 $hebrewDay -= $days;
1862                                                 # Try Adar II
1863                                                 $days = 29;
1864                                                 if ( $hebrewDay <= $days ) {
1865                                                         # Day in Adar II
1866                                                         $hebrewMonth = 14;
1867                                                 }
1868                                         }
1869                                 }
1870                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1871                                 # Cheshvan in a complete year (otherwise as the rule below)
1872                                 $days = 30;
1873                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1874                                 # Kislev in an incomplete year (otherwise as the rule below)
1875                                 $days = 29;
1876                         } else {
1877                                 # Odd months have 30 days, even have 29
1878                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1879                         }
1880                         if ( $hebrewDay <= $days ) {
1881                                 # In the current month
1882                                 break;
1883                         } else {
1884                                 # Subtract the days of the current month
1885                                 $hebrewDay -= $days;
1886                                 # Try in the next month
1887                                 $hebrewMonth++;
1888                         }
1889                 }
1890
1891                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1892         }
1893
1894         /**
1895          * This calculates the Hebrew year start, as days since 1 September.
1896          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1897          * Used for Hebrew date.
1898          *
1899          * @param int $year
1900          *
1901          * @return string
1902          */
1903         private static function hebrewYearStart( $year ) {
1904                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1905                 $b = intval( ( $year - 1 ) % 4 );
1906                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1907                 if ( $m < 0 ) {
1908                         $m--;
1909                 }
1910                 $Mar = intval( $m );
1911                 if ( $m < 0 ) {
1912                         $m++;
1913                 }
1914                 $m -= $Mar;
1915
1916                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1917                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1918                         $Mar++;
1919                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1920                         $Mar += 2;
1921                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1922                         $Mar++;
1923                 }
1924
1925                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1926                 return $Mar;
1927         }
1928
1929         /**
1930          * Algorithm to convert Gregorian dates to Thai solar dates,
1931          * Minguo dates or Minguo dates.
1932          *
1933          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1934          *       https://en.wikipedia.org/wiki/Minguo_calendar
1935          *       https://en.wikipedia.org/wiki/Japanese_era_name
1936          *
1937          * @param string $ts 14-character timestamp
1938          * @param string $cName Calender name
1939          * @return array Converted year, month, day
1940          */
1941         private static function tsToYear( $ts, $cName ) {
1942                 $gy = substr( $ts, 0, 4 );
1943                 $gm = substr( $ts, 4, 2 );
1944                 $gd = substr( $ts, 6, 2 );
1945
1946                 if ( !strcmp( $cName, 'thai' ) ) {
1947                         # Thai solar dates
1948                         # Add 543 years to the Gregorian calendar
1949                         # Months and days are identical
1950                         $gy_offset = $gy + 543;
1951                         # fix for dates between 1912 and 1941
1952                         # https://en.wikipedia.org/?oldid=836596673#New_year
1953                         if ( $gy >= 1912 && $gy <= 1940 ) {
1954                                 if ( $gm <= 3 ) {
1955                                         $gy_offset--;
1956                                 }
1957                                 $gm = ( $gm - 3 ) % 12;
1958                         }
1959                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1960                         # Minguo dates
1961                         # Deduct 1911 years from the Gregorian calendar
1962                         # Months and days are identical
1963                         $gy_offset = $gy - 1911;
1964                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1965                         # Nengō dates up to Meiji period
1966                         # Deduct years from the Gregorian calendar
1967                         # depending on the nengo periods
1968                         # Months and days are identical
1969                         if ( ( $gy < 1912 )
1970                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1971                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1972                         ) {
1973                                 # Meiji period
1974                                 $gy_gannen = $gy - 1868 + 1;
1975                                 $gy_offset = $gy_gannen;
1976                                 if ( $gy_gannen == 1 ) {
1977                                         $gy_offset = '元';
1978                                 }
1979                                 $gy_offset = '明治' . $gy_offset;
1980                         } elseif (
1981                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1982                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1983                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1984                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1985                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1986                         ) {
1987                                 # Taishō period
1988                                 $gy_gannen = $gy - 1912 + 1;
1989                                 $gy_offset = $gy_gannen;
1990                                 if ( $gy_gannen == 1 ) {
1991                                         $gy_offset = '元';
1992                                 }
1993                                 $gy_offset = '大正' . $gy_offset;
1994                         } elseif (
1995                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1996                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1997                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1998                         ) {
1999                                 # Shōwa period
2000                                 $gy_gannen = $gy - 1926 + 1;
2001                                 $gy_offset = $gy_gannen;
2002                                 if ( $gy_gannen == 1 ) {
2003                                         $gy_offset = '元';
2004                                 }
2005                                 $gy_offset = '昭和' . $gy_offset;
2006                         } else {
2007                                 # Heisei period
2008                                 $gy_gannen = $gy - 1989 + 1;
2009                                 $gy_offset = $gy_gannen;
2010                                 if ( $gy_gannen == 1 ) {
2011                                         $gy_offset = '元';
2012                                 }
2013                                 $gy_offset = '平成' . $gy_offset;
2014                         }
2015                 } else {
2016                         $gy_offset = $gy;
2017                 }
2018
2019                 return [ $gy_offset, $gm, $gd ];
2020         }
2021
2022         /**
2023          * Gets directionality of the first strongly directional codepoint, for embedBidi()
2024          *
2025          * This is the rule the BIDI algorithm uses to determine the directionality of
2026          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2027          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2028          *
2029          * TODO: Does not handle BIDI control characters inside the text.
2030          * TODO: Does not handle unallocated characters.
2031          *
2032          * @param string $text Text to test
2033          * @return null|string Directionality ('ltr' or 'rtl') or null
2034          */
2035         private static function strongDirFromContent( $text = '' ) {
2036                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2037                         return null;
2038                 }
2039                 if ( $matches[1] === '' ) {
2040                         return 'rtl';
2041                 }
2042                 return 'ltr';
2043         }
2044
2045         /**
2046          * Roman number formatting up to 10000
2047          *
2048          * @param int $num
2049          *
2050          * @return string
2051          */
2052         static function romanNumeral( $num ) {
2053                 static $table = [
2054                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2055                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2056                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2057                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2058                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2059                 ];
2060
2061                 $num = intval( $num );
2062                 if ( $num > 10000 || $num <= 0 ) {
2063                         return $num;
2064                 }
2065
2066                 $s = '';
2067                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2068                         if ( $num >= $pow10 ) {
2069                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2070                         }
2071                         $num = $num % $pow10;
2072                 }
2073                 return $s;
2074         }
2075
2076         /**
2077          * Hebrew Gematria number formatting up to 9999
2078          *
2079          * @param int $num
2080          *
2081          * @return string
2082          */
2083         static function hebrewNumeral( $num ) {
2084                 static $table = [
2085                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2086                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2087                         [ '',
2088                                 [ 'ק' ],
2089                                 [ 'ר' ],
2090                                 [ 'ש' ],
2091                                 [ 'ת' ],
2092                                 [ 'ת', 'ק' ],
2093                                 [ 'ת', 'ר' ],
2094                                 [ 'ת', 'ש' ],
2095                                 [ 'ת', 'ת' ],
2096                                 [ 'ת', 'ת', 'ק' ],
2097                                 [ 'ת', 'ת', 'ר' ],
2098                         ],
2099                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2100                 ];
2101
2102                 $num = intval( $num );
2103                 if ( $num > 9999 || $num <= 0 ) {
2104                         return $num;
2105                 }
2106
2107                 // Round thousands have special notations
2108                 if ( $num === 1000 ) {
2109                         return "א' אלף";
2110                 } elseif ( $num % 1000 === 0 ) {
2111                         return $table[0][$num / 1000] . "' אלפים";
2112                 }
2113
2114                 $letters = [];
2115
2116                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2117                         if ( $num >= $pow10 ) {
2118                                 if ( $num === 15 || $num === 16 ) {
2119                                         $letters[] = $table[0][9];
2120                                         $letters[] = $table[0][$num - 9];
2121                                         $num = 0;
2122                                 } else {
2123                                         $letters = array_merge(
2124                                                 $letters,
2125                                                 (array)$table[$i][intval( $num / $pow10 )]
2126                                         );
2127
2128                                         if ( $pow10 === 1000 ) {
2129                                                 $letters[] = "'";
2130                                         }
2131                                 }
2132                         }
2133
2134                         $num = $num % $pow10;
2135                 }
2136
2137                 $preTransformLength = count( $letters );
2138                 if ( $preTransformLength === 1 ) {
2139                         // Add geresh (single quote) to one-letter numbers
2140                         $letters[] = "'";
2141                 } else {
2142                         $lastIndex = $preTransformLength - 1;
2143                         $letters[$lastIndex] = str_replace(
2144                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2145                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2146                                 $letters[$lastIndex]
2147                         );
2148
2149                         // Add gershayim (double quote) to multiple-letter numbers,
2150                         // but exclude numbers with only one letter after the thousands
2151                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2152                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2153                                 $letters[] = "'";
2154                         } else {
2155                                 array_splice( $letters, -1, 0, '"' );
2156                         }
2157                 }
2158
2159                 return implode( $letters );
2160         }
2161
2162         /**
2163          * Used by date() and time() to adjust the time output.
2164          *
2165          * @param string $ts The time in date('YmdHis') format
2166          * @param mixed $tz Adjust the time by this amount (default false, mean we
2167          *   get user timecorrection setting)
2168          * @return int
2169          */
2170         public function userAdjust( $ts, $tz = false ) {
2171                 global $wgUser, $wgLocalTZoffset;
2172
2173                 if ( $tz === false ) {
2174                         $tz = $wgUser->getOption( 'timecorrection' );
2175                 }
2176
2177                 $data = explode( '|', $tz, 3 );
2178
2179                 if ( $data[0] == 'ZoneInfo' ) {
2180                         try {
2181                                 $userTZ = new DateTimeZone( $data[2] );
2182                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2183                                 $date->setTimezone( $userTZ );
2184                                 return $date->format( 'YmdHis' );
2185                         } catch ( Exception $e ) {
2186                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2187                                 $data[0] = 'Offset';
2188                         }
2189                 }
2190
2191                 if ( $data[0] == 'System' || $tz == '' ) {
2192                         # Global offset in minutes.
2193                         $minDiff = $wgLocalTZoffset;
2194                 } elseif ( $data[0] == 'Offset' ) {
2195                         $minDiff = intval( $data[1] );
2196                 } else {
2197                         $data = explode( ':', $tz );
2198                         if ( count( $data ) == 2 ) {
2199                                 $data[0] = intval( $data[0] );
2200                                 $data[1] = intval( $data[1] );
2201                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2202                                 if ( $data[0] < 0 ) {
2203                                         $minDiff = -$minDiff;
2204                                 }
2205                         } else {
2206                                 $minDiff = intval( $data[0] ) * 60;
2207                         }
2208                 }
2209
2210                 # No difference ? Return time unchanged
2211                 if ( 0 == $minDiff ) {
2212                         return $ts;
2213                 }
2214
2215                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2216                 # Generate an adjusted date; take advantage of the fact that mktime
2217                 # will normalize out-of-range values so we don't have to split $minDiff
2218                 # into hours and minutes.
2219                 $t = mktime( (
2220                         (int)substr( $ts, 8, 2 ) ), # Hours
2221                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2222                         (int)substr( $ts, 12, 2 ), # Seconds
2223                         (int)substr( $ts, 4, 2 ), # Month
2224                         (int)substr( $ts, 6, 2 ), # Day
2225                         (int)substr( $ts, 0, 4 ) ); # Year
2226
2227                 $date = date( 'YmdHis', $t );
2228                 Wikimedia\restoreWarnings();
2229
2230                 return $date;
2231         }
2232
2233         /**
2234          * This is meant to be used by time(), date(), and timeanddate() to get
2235          * the date preference they're supposed to use, it should be used in
2236          * all children.
2237          *
2238          *     function timeanddate([...], $format = true) {
2239          *       $datePreference = $this->dateFormat($format);
2240          *       [...]
2241          *     }
2242          *
2243          * @param int|string|bool $usePrefs If true, the user's preference is used
2244          *   if false, the site/language default is used
2245          *   if int/string, assumed to be a format.
2246          * @return string
2247          */
2248         function dateFormat( $usePrefs = true ) {
2249                 global $wgUser;
2250
2251                 if ( is_bool( $usePrefs ) ) {
2252                         if ( $usePrefs ) {
2253                                 $datePreference = $wgUser->getDatePreference();
2254                         } else {
2255                                 $datePreference = (string)User::getDefaultOption( 'date' );
2256                         }
2257                 } else {
2258                         $datePreference = (string)$usePrefs;
2259                 }
2260
2261                 // return int
2262                 if ( $datePreference == '' ) {
2263                         return 'default';
2264                 }
2265
2266                 return $datePreference;
2267         }
2268
2269         /**
2270          * Get a format string for a given type and preference
2271          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2272          * @param string $pref The format name as it appears in Messages*.php under
2273          *  $datePreferences.
2274          *
2275          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2276          *
2277          * @return string
2278          */
2279         function getDateFormatString( $type, $pref ) {
2280                 $wasDefault = false;
2281                 if ( $pref == 'default' ) {
2282                         $wasDefault = true;
2283                         $pref = $this->getDefaultDateFormat();
2284                 }
2285
2286                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2287                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2288
2289                         if ( $type === 'pretty' && $df === null ) {
2290                                 $df = $this->getDateFormatString( 'date', $pref );
2291                         }
2292
2293                         if ( !$wasDefault && $df === null ) {
2294                                 $pref = $this->getDefaultDateFormat();
2295                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2296                         }
2297
2298                         $this->dateFormatStrings[$type][$pref] = $df;
2299                 }
2300                 return $this->dateFormatStrings[$type][$pref];
2301         }
2302
2303         /**
2304          * @param string $ts The time format which needs to be turned into a
2305          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2306          * @param bool $adj Whether to adjust the time output according to the
2307          *   user configured offset ($timecorrection)
2308          * @param mixed $format True to use user's date format preference
2309          * @param string|bool $timecorrection The time offset as returned by
2310          *   validateTimeZone() in Special:Preferences
2311          * @return string
2312          */
2313         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2314                 $ts = wfTimestamp( TS_MW, $ts );
2315                 if ( $adj ) {
2316                         $ts = $this->userAdjust( $ts, $timecorrection );
2317                 }
2318                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2319                 return $this->sprintfDate( $df, $ts );
2320         }
2321
2322         /**
2323          * @param string $ts The time format which needs to be turned into a
2324          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2325          * @param bool $adj Whether to adjust the time output according to the
2326          *   user configured offset ($timecorrection)
2327          * @param mixed $format True to use user's date format preference
2328          * @param string|bool $timecorrection The time offset as returned by
2329          *   validateTimeZone() in Special:Preferences
2330          * @return string
2331          */
2332         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2333                 $ts = wfTimestamp( TS_MW, $ts );
2334                 if ( $adj ) {
2335                         $ts = $this->userAdjust( $ts, $timecorrection );
2336                 }
2337                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2338                 return $this->sprintfDate( $df, $ts );
2339         }
2340
2341         /**
2342          * @param string $ts The time format which needs to be turned into a
2343          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2344          * @param bool $adj Whether to adjust the time output according to the
2345          *   user configured offset ($timecorrection)
2346          * @param mixed $format What format to return, if it's false output the
2347          *   default one (default true)
2348          * @param string|bool $timecorrection The time offset as returned by
2349          *   validateTimeZone() in Special:Preferences
2350          * @return string
2351          */
2352         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2353                 $ts = wfTimestamp( TS_MW, $ts );
2354                 if ( $adj ) {
2355                         $ts = $this->userAdjust( $ts, $timecorrection );
2356                 }
2357                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2358                 return $this->sprintfDate( $df, $ts );
2359         }
2360
2361         /**
2362          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2363          *
2364          * @since 1.20
2365          *
2366          * @param int $seconds The amount of seconds.
2367          * @param array $chosenIntervals The intervals to enable.
2368          *
2369          * @return string
2370          */
2371         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2372                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2373
2374                 $segments = [];
2375
2376                 foreach ( $intervals as $intervalName => $intervalValue ) {
2377                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2378                         // duration-years, duration-decades, duration-centuries, duration-millennia
2379                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2380                         $segments[] = $message->inLanguage( $this )->escaped();
2381                 }
2382
2383                 return $this->listToText( $segments );
2384         }
2385
2386         /**
2387          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2388          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2389          *
2390          * @since 1.20
2391          *
2392          * @param int $seconds The amount of seconds.
2393          * @param array $chosenIntervals The intervals to enable.
2394          *
2395          * @return array
2396          */
2397         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2398                 if ( empty( $chosenIntervals ) ) {
2399                         $chosenIntervals = [
2400                                 'millennia',
2401                                 'centuries',
2402                                 'decades',
2403                                 'years',
2404                                 'days',
2405                                 'hours',
2406                                 'minutes',
2407                                 'seconds'
2408                         ];
2409                 }
2410
2411                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2412                 $sortedNames = array_keys( $intervals );
2413                 $smallestInterval = array_pop( $sortedNames );
2414
2415                 $segments = [];
2416
2417                 foreach ( $intervals as $name => $length ) {
2418                         $value = floor( $seconds / $length );
2419
2420                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2421                                 $seconds -= $value * $length;
2422                                 $segments[$name] = $value;
2423                         }
2424                 }
2425
2426                 return $segments;
2427         }
2428
2429         /**
2430          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2431          *
2432          * @param string $type Can be 'date', 'time' or 'both'
2433          * @param string $ts The time format which needs to be turned into a
2434          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2435          * @param User $user User object used to get preferences for timezone and format
2436          * @param array $options Array, can contain the following keys:
2437          *   - 'timecorrection': time correction, can have the following values:
2438          *     - true: use user's preference
2439          *     - false: don't use time correction
2440          *     - int: value of time correction in minutes
2441          *   - 'format': format to use, can have the following values:
2442          *     - true: use user's preference
2443          *     - false: use default preference
2444          *     - string: format to use
2445          * @since 1.19
2446          * @return string
2447          */
2448         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2449                 $ts = wfTimestamp( TS_MW, $ts );
2450                 $options += [ 'timecorrection' => true, 'format' => true ];
2451                 if ( $options['timecorrection'] !== false ) {
2452                         if ( $options['timecorrection'] === true ) {
2453                                 $offset = $user->getOption( 'timecorrection' );
2454                         } else {
2455                                 $offset = $options['timecorrection'];
2456                         }
2457                         $ts = $this->userAdjust( $ts, $offset );
2458                 }
2459                 if ( $options['format'] === true ) {
2460                         $format = $user->getDatePreference();
2461                 } else {
2462                         $format = $options['format'];
2463                 }
2464                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2465                 return $this->sprintfDate( $df, $ts );
2466         }
2467
2468         /**
2469          * Get the formatted date for the given timestamp and formatted for
2470          * the given user.
2471          *
2472          * @param mixed $ts Mixed: the time format which needs to be turned into a
2473          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2474          * @param User $user User object used to get preferences for timezone and format
2475          * @param array $options Array, can contain the following keys:
2476          *   - 'timecorrection': time correction, can have the following values:
2477          *     - true: use user's preference
2478          *     - false: don't use time correction
2479          *     - int: value of time correction in minutes
2480          *   - 'format': format to use, can have the following values:
2481          *     - true: use user's preference
2482          *     - false: use default preference
2483          *     - string: format to use
2484          * @since 1.19
2485          * @return string
2486          */
2487         public function userDate( $ts, User $user, array $options = [] ) {
2488                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2489         }
2490
2491         /**
2492          * Get the formatted time for the given timestamp and formatted for
2493          * the given user.
2494          *
2495          * @param mixed $ts The time format which needs to be turned into a
2496          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2497          * @param User $user User object used to get preferences for timezone and format
2498          * @param array $options Array, can contain the following keys:
2499          *   - 'timecorrection': time correction, can have the following values:
2500          *     - true: use user's preference
2501          *     - false: don't use time correction
2502          *     - int: value of time correction in minutes
2503          *   - 'format': format to use, can have the following values:
2504          *     - true: use user's preference
2505          *     - false: use default preference
2506          *     - string: format to use
2507          * @since 1.19
2508          * @return string
2509          */
2510         public function userTime( $ts, User $user, array $options = [] ) {
2511                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2512         }
2513
2514         /**
2515          * Get the formatted date and time for the given timestamp and formatted for
2516          * the given user.
2517          *
2518          * @param mixed $ts The time format which needs to be turned into a
2519          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2520          * @param User $user User object used to get preferences for timezone and format
2521          * @param array $options Array, can contain the following keys:
2522          *   - 'timecorrection': time correction, can have the following values:
2523          *     - true: use user's preference
2524          *     - false: don't use time correction
2525          *     - int: value of time correction in minutes
2526          *   - 'format': format to use, can have the following values:
2527          *     - true: use user's preference
2528          *     - false: use default preference
2529          *     - string: format to use
2530          * @since 1.19
2531          * @return string
2532          */
2533         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2534                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2535         }
2536
2537         /**
2538          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2539          *
2540          * Determine the difference between the timestamp and the current time, and
2541          * generate a readable timestamp by returning "<N> <units> ago", where the
2542          * largest possible unit is used.
2543          *
2544          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2545          *
2546          * @param MWTimestamp $time
2547          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2548          * @param User|null $user User the timestamp is being generated for
2549          *  (or null to use main context's user)
2550          * @return string Formatted timestamp
2551          */
2552         public function getHumanTimestamp(
2553                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2554         ) {
2555                 if ( $relativeTo === null ) {
2556                         $relativeTo = new MWTimestamp();
2557                 }
2558                 if ( $user === null ) {
2559                         $user = RequestContext::getMain()->getUser();
2560                 }
2561
2562                 // Adjust for the user's timezone.
2563                 $offsetThis = $time->offsetForUser( $user );
2564                 $offsetRel = $relativeTo->offsetForUser( $user );
2565
2566                 $ts = '';
2567                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2568                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2569                 }
2570
2571                 // Reset the timezone on the objects.
2572                 $time->timestamp->sub( $offsetThis );
2573                 $relativeTo->timestamp->sub( $offsetRel );
2574
2575                 return $ts;
2576         }
2577
2578         /**
2579          * Convert an MWTimestamp into a pretty human-readable timestamp using
2580          * the given user preferences and relative base time.
2581          *
2582          * @see Language::getHumanTimestamp
2583          * @param MWTimestamp $ts Timestamp to prettify
2584          * @param MWTimestamp $relativeTo Base timestamp
2585          * @param User $user User preferences to use
2586          * @return string Human timestamp
2587          * @since 1.26
2588          */
2589         private function getHumanTimestampInternal(
2590                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2591         ) {
2592                 $diff = $ts->diff( $relativeTo );
2593                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2594                         (int)$relativeTo->timestamp->format( 'w' ) );
2595                 $days = $diff->days ?: (int)$diffDay;
2596                 if ( $diff->invert || $days > 5
2597                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2598                 ) {
2599                         // Timestamps are in different years: use full timestamp
2600                         // Also do full timestamp for future dates
2601                         /**
2602                          * @todo FIXME: Add better handling of future timestamps.
2603                          */
2604                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2605                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2606                 } elseif ( $days > 5 ) {
2607                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2608                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2609                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2610                 } elseif ( $days > 1 ) {
2611                         // Timestamp within the past week: show the day of the week and time
2612                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2613                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2614                         // Messages:
2615                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2616                         $ts = wfMessage( "$weekday-at" )
2617                                 ->inLanguage( $this )
2618                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2619                                 ->text();
2620                 } elseif ( $days == 1 ) {
2621                         // Timestamp was yesterday: say 'yesterday' and the time.
2622                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2623                         $ts = wfMessage( 'yesterday-at' )
2624                                 ->inLanguage( $this )
2625                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2626                                 ->text();
2627                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2628                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2629                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2630                         $ts = wfMessage( 'today-at' )
2631                                 ->inLanguage( $this )
2632                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2633                                 ->text();
2634
2635                 // From here on in, the timestamp was soon enough ago so that we can simply say
2636                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2637                 } elseif ( $diff->h == 1 ) {
2638                         // Less than 90 minutes, but more than an hour ago.
2639                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2640                 } elseif ( $diff->i >= 1 ) {
2641                         // A few minutes ago.
2642                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2643                 } elseif ( $diff->s >= 30 ) {
2644                         // Less than a minute, but more than 30 sec ago.
2645                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2646                 } else {
2647                         // Less than 30 seconds ago.
2648                         $ts = wfMessage( 'just-now' )->text();
2649                 }
2650
2651                 return $ts;
2652         }
2653
2654         /**
2655          * @param string $key
2656          * @return string|null
2657          */
2658         public function getMessage( $key ) {
2659                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2660         }
2661
2662         /**
2663          * @return array
2664          */
2665         function getAllMessages() {
2666                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2667         }
2668
2669         /**
2670          * @param string $in
2671          * @param string $out
2672          * @param string $string
2673          * @return string
2674          */
2675         public function iconv( $in, $out, $string ) {
2676                 # Even with //IGNORE iconv can whine about illegal characters in
2677                 # *input* string. We just ignore those too.
2678                 # REF: https://bugs.php.net/bug.php?id=37166
2679                 # REF: https://phabricator.wikimedia.org/T18885
2680                 Wikimedia\suppressWarnings();
2681                 $text = iconv( $in, $out . '//IGNORE', $string );
2682                 Wikimedia\restoreWarnings();
2683                 return $text;
2684         }
2685
2686         // callback functions for ucwords(), ucwordbreaks()
2687
2688         /**
2689          * @param array $matches
2690          * @return mixed|string
2691          */
2692         function ucwordbreaksCallbackAscii( $matches ) {
2693                 return $this->ucfirst( $matches[1] );
2694         }
2695
2696         /**
2697          * @param array $matches
2698          * @return string
2699          */
2700         function ucwordbreaksCallbackMB( $matches ) {
2701                 return mb_strtoupper( $matches[0] );
2702         }
2703
2704         /**
2705          * @param array $matches
2706          * @return string
2707          */
2708         function ucwordsCallbackMB( $matches ) {
2709                 return mb_strtoupper( $matches[0] );
2710         }
2711
2712         /**
2713          * Make a string's first character uppercase
2714          *
2715          * @param string $str
2716          *
2717          * @return string
2718          */
2719         public function ucfirst( $str ) {
2720                 $o = ord( $str );
2721                 if ( $o < 96 ) { // if already uppercase...
2722                         return $str;
2723                 } elseif ( $o < 128 ) {
2724                         return ucfirst( $str ); // use PHP's ucfirst()
2725                 } else {
2726                         // fall back to more complex logic in case of multibyte strings
2727                         return $this->uc( $str, true );
2728                 }
2729         }
2730
2731         /**
2732          * Convert a string to uppercase
2733          *
2734          * @param string $str
2735          * @param bool $first
2736          *
2737          * @return string
2738          */
2739         public function uc( $str, $first = false ) {
2740                 if ( $first ) {
2741                         if ( $this->isMultibyte( $str ) ) {
2742                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2743                         } else {
2744                                 return ucfirst( $str );
2745                         }
2746                 } else {
2747                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2748                 }
2749         }
2750
2751         /**
2752          * @param string $str
2753          * @return mixed|string
2754          */
2755         function lcfirst( $str ) {
2756                 $o = ord( $str );
2757                 if ( !$o ) {
2758                         return strval( $str );
2759                 } elseif ( $o >= 128 ) {
2760                         return $this->lc( $str, true );
2761                 } elseif ( $o > 96 ) {
2762                         return $str;
2763                 } else {
2764                         $str[0] = strtolower( $str[0] );
2765                         return $str;
2766                 }
2767         }
2768
2769         /**
2770          * @param string $str
2771          * @param bool $first
2772          * @return mixed|string
2773          */
2774         function lc( $str, $first = false ) {
2775                 if ( $first ) {
2776                         if ( $this->isMultibyte( $str ) ) {
2777                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2778                         } else {
2779                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2780                         }
2781                 } else {
2782                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2783                 }
2784         }
2785
2786         /**
2787          * @param string $str
2788          * @return bool
2789          */
2790         function isMultibyte( $str ) {
2791                 return strlen( $str ) !== mb_strlen( $str );
2792         }
2793
2794         /**
2795          * @param string $str
2796          * @return mixed|string
2797          */
2798         function ucwords( $str ) {
2799                 if ( $this->isMultibyte( $str ) ) {
2800                         $str = $this->lc( $str );
2801
2802                         // regexp to find first letter in each word (i.e. after each space)
2803                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2804
2805                         // function to use to capitalize a single char
2806                         return preg_replace_callback(
2807                                 $replaceRegexp,
2808                                 [ $this, 'ucwordsCallbackMB' ],
2809                                 $str
2810                         );
2811                 } else {
2812                         return ucwords( strtolower( $str ) );
2813                 }
2814         }
2815
2816         /**
2817          * capitalize words at word breaks
2818          *
2819          * @param string $str
2820          * @return mixed
2821          */
2822         function ucwordbreaks( $str ) {
2823                 if ( $this->isMultibyte( $str ) ) {
2824                         $str = $this->lc( $str );
2825
2826                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2827                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2828
2829                         // find first letter after word break
2830                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2831                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2832
2833                         return preg_replace_callback(
2834                                 $replaceRegexp,
2835                                 [ $this, 'ucwordbreaksCallbackMB' ],
2836                                 $str
2837                         );
2838                 } else {
2839                         return preg_replace_callback(
2840                                 '/\b([\w\x80-\xff]+)\b/',
2841                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2842                                 $str
2843                         );
2844                 }
2845         }
2846
2847         /**
2848          * Return a case-folded representation of $s
2849          *
2850          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2851          * and $s2 are the same except for the case of their characters. It is not
2852          * necessary for the value returned to make sense when displayed.
2853          *
2854          * Do *not* perform any other normalisation in this function. If a caller
2855          * uses this function when it should be using a more general normalisation
2856          * function, then fix the caller.
2857          *
2858          * @param string $s
2859          *
2860          * @return string
2861          */
2862         function caseFold( $s ) {
2863                 return $this->uc( $s );
2864         }
2865
2866         /**
2867          * @param string $s
2868          * @return string
2869          * @throws MWException
2870          */
2871         function checkTitleEncoding( $s ) {
2872                 if ( is_array( $s ) ) {
2873                         throw new MWException( 'Given array to checkTitleEncoding.' );
2874                 }
2875                 if ( StringUtils::isUtf8( $s ) ) {
2876                         return $s;
2877                 }
2878
2879                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2880         }
2881
2882         /**
2883          * @return array
2884          */
2885         function fallback8bitEncoding() {
2886                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2887         }
2888
2889         /**
2890          * Most writing systems use whitespace to break up words.
2891          * Some languages such as Chinese don't conventionally do this,
2892          * which requires special handling when breaking up words for
2893          * searching etc.
2894          *
2895          * @return bool
2896          */
2897         function hasWordBreaks() {
2898                 return true;
2899         }
2900
2901         /**
2902          * Some languages such as Chinese require word segmentation,
2903          * Specify such segmentation when overridden in derived class.
2904          *
2905          * @param string $string
2906          * @return string
2907          */
2908         function segmentByWord( $string ) {
2909                 return $string;
2910         }
2911
2912         /**
2913          * Some languages have special punctuation need to be normalized.
2914          * Make such changes here.
2915          *
2916          * @param string $string
2917          * @return string
2918          */
2919         function normalizeForSearch( $string ) {
2920                 return self::convertDoubleWidth( $string );
2921         }
2922
2923         /**
2924          * convert double-width roman characters to single-width.
2925          * range: ff00-ff5f ~= 0020-007f
2926          *
2927          * @param string $string
2928          *
2929          * @return string
2930          */
2931         protected static function convertDoubleWidth( $string ) {
2932                 static $full = null;
2933                 static $half = null;
2934
2935                 if ( $full === null ) {
2936                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2937                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2938                         $full = str_split( $fullWidth, 3 );
2939                         $half = str_split( $halfWidth );
2940                 }
2941
2942                 $string = str_replace( $full, $half, $string );
2943                 return $string;
2944         }
2945
2946         /**
2947          * @param string $string
2948          * @param string $pattern
2949          * @return string
2950          */
2951         protected static function insertSpace( $string, $pattern ) {
2952                 $string = preg_replace( $pattern, " $1 ", $string );
2953                 $string = preg_replace( '/ +/', ' ', $string );
2954                 return $string;
2955         }
2956
2957         /**
2958          * @param array $termsArray
2959          * @return array
2960          */
2961         function convertForSearchResult( $termsArray ) {
2962                 # some languages, e.g. Chinese, need to do a conversion
2963                 # in order for search results to be displayed correctly
2964                 return $termsArray;
2965         }
2966
2967         /**
2968          * Get the first character of a string.
2969          *
2970          * @param string $s
2971          * @return string
2972          */
2973         function firstChar( $s ) {
2974                 $matches = [];
2975                 preg_match(
2976                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2977                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2978                         $s,
2979                         $matches
2980                 );
2981
2982                 if ( isset( $matches[1] ) ) {
2983                         if ( strlen( $matches[1] ) != 3 ) {
2984                                 return $matches[1];
2985                         }
2986
2987                         // Break down Hangul syllables to grab the first jamo
2988                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2989                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2990                                 return $matches[1];
2991                         } elseif ( $code < 0xb098 ) {
2992                                 return "\u{3131}";
2993                         } elseif ( $code < 0xb2e4 ) {
2994                                 return "\u{3134}";
2995                         } elseif ( $code < 0xb77c ) {
2996                                 return "\u{3137}";
2997                         } elseif ( $code < 0xb9c8 ) {
2998                                 return "\u{3139}";
2999                         } elseif ( $code < 0xbc14 ) {
3000                                 return "\u{3141}";
3001                         } elseif ( $code < 0xc0ac ) {
3002                                 return "\u{3142}";
3003                         } elseif ( $code < 0xc544 ) {
3004                                 return "\u{3145}";
3005                         } elseif ( $code < 0xc790 ) {
3006                                 return "\u{3147}";
3007                         } elseif ( $code < 0xcc28 ) {
3008                                 return "\u{3148}";
3009                         } elseif ( $code < 0xce74 ) {
3010                                 return "\u{314A}";
3011                         } elseif ( $code < 0xd0c0 ) {
3012                                 return "\u{314B}";
3013                         } elseif ( $code < 0xd30c ) {
3014                                 return "\u{314C}";
3015                         } elseif ( $code < 0xd558 ) {
3016                                 return "\u{314D}";
3017                         } else {
3018                                 return "\u{314E}";
3019                         }
3020                 } else {
3021                         return '';
3022                 }
3023         }
3024
3025         /**
3026          * @deprecated No-op since 1.28
3027          */
3028         function initEncoding() {
3029                 wfDeprecated( __METHOD__, '1.28' );
3030                 // No-op.
3031         }
3032
3033         /**
3034          * @param string $s
3035          * @return string
3036          * @deprecated No-op since 1.28
3037          */
3038         function recodeForEdit( $s ) {
3039                 wfDeprecated( __METHOD__, '1.28' );
3040                 return $s;
3041         }
3042
3043         /**
3044          * @param string $s
3045          * @return string
3046          * @deprecated No-op since 1.28
3047          */
3048         function recodeInput( $s ) {
3049                 wfDeprecated( __METHOD__, '1.28' );
3050                 return $s;
3051         }
3052
3053         /**
3054          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3055          * also cleans up certain backwards-compatible sequences, converting them
3056          * to the modern Unicode equivalent.
3057          *
3058          * This is language-specific for performance reasons only.
3059          *
3060          * @param string $s
3061          *
3062          * @return string
3063          */
3064         public function normalize( $s ) {
3065                 global $wgAllUnicodeFixes;
3066                 $s = UtfNormal\Validator::cleanUp( $s );
3067                 if ( $wgAllUnicodeFixes ) {
3068                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3069                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3070                 }
3071
3072                 return $s;
3073         }
3074
3075         /**
3076          * Transform a string using serialized data stored in the given file (which
3077          * must be in the serialized subdirectory of $IP). The file contains pairs
3078          * mapping source characters to destination characters.
3079          *
3080          * The data is cached in process memory. This will go faster if you have the
3081          * FastStringSearch extension.
3082          *
3083          * @param string $file
3084          * @param string $string
3085          *
3086          * @throws MWException
3087          * @return string
3088          */
3089         protected function transformUsingPairFile( $file, $string ) {
3090                 if ( !isset( $this->transformData[$file] ) ) {
3091                         global $IP;
3092                         $data = require "$IP/languages/data/{$file}";
3093                         $this->transformData[$file] = new ReplacementArray( $data );
3094                 }
3095                 return $this->transformData[$file]->replace( $string );
3096         }
3097
3098         /**
3099          * For right-to-left language support
3100          *
3101          * @return bool
3102          */
3103         function isRTL() {
3104                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3105         }
3106
3107         /**
3108          * Return the correct HTML 'dir' attribute value for this language.
3109          * @return string
3110          */
3111         function getDir() {
3112                 return $this->isRTL() ? 'rtl' : 'ltr';
3113         }
3114
3115         /**
3116          * Return 'left' or 'right' as appropriate alignment for line-start
3117          * for this language's text direction.
3118          *
3119          * Should be equivalent to CSS3 'start' text-align value....
3120          *
3121          * @return string
3122          */
3123         function alignStart() {
3124                 return $this->isRTL() ? 'right' : 'left';
3125         }
3126
3127         /**
3128          * Return 'right' or 'left' as appropriate alignment for line-end
3129          * for this language's text direction.
3130          *
3131          * Should be equivalent to CSS3 'end' text-align value....
3132          *
3133          * @return string
3134          */
3135         function alignEnd() {
3136                 return $this->isRTL() ? 'left' : 'right';
3137         }
3138
3139         /**
3140          * A hidden direction mark (LRM or RLM), depending on the language direction.
3141          * Unlike getDirMark(), this function returns the character as an HTML entity.
3142          * This function should be used when the output is guaranteed to be HTML,
3143          * because it makes the output HTML source code more readable. When
3144          * the output is plain text or can be escaped, getDirMark() should be used.
3145          *
3146          * @param bool $opposite Get the direction mark opposite to your language
3147          * @return string
3148          * @since 1.20
3149          */
3150         function getDirMarkEntity( $opposite = false ) {
3151                 if ( $opposite ) {
3152                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3153                 }
3154                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3155         }
3156
3157         /**
3158          * A hidden direction mark (LRM or RLM), depending on the language direction.
3159          * This function produces them as invisible Unicode characters and
3160          * the output may be hard to read and debug, so it should only be used
3161          * when the output is plain text or can be escaped. When the output is
3162          * HTML, use getDirMarkEntity() instead.
3163          *
3164          * @param bool $opposite Get the direction mark opposite to your language
3165          * @return string
3166          */
3167         function getDirMark( $opposite = false ) {
3168                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3169                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3170                 if ( $opposite ) {
3171                         return $this->isRTL() ? $lrm : $rlm;
3172                 }
3173                 return $this->isRTL() ? $rlm : $lrm;
3174         }
3175
3176         /**
3177          * @return array
3178          */
3179         function capitalizeAllNouns() {
3180                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3181         }
3182
3183         /**
3184          * An arrow, depending on the language direction.
3185          *
3186          * @param string $direction The direction of the arrow: forwards (default),
3187          *   backwards, left, right, up, down.
3188          * @return string
3189          */
3190         function getArrow( $direction = 'forwards' ) {
3191                 switch ( $direction ) {
3192                         case 'forwards':
3193                                 return $this->isRTL() ? '←' : '→';
3194                         case 'backwards':
3195                                 return $this->isRTL() ? '→' : '←';
3196                         case 'left':
3197                                 return '←';
3198                         case 'right':
3199                                 return '→';
3200                         case 'up':
3201                                 return '↑';
3202                         case 'down':
3203                                 return '↓';
3204                 }
3205         }
3206
3207         /**
3208          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3209          *
3210          * @return bool
3211          */
3212         function linkPrefixExtension() {
3213                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3214         }
3215
3216         /**
3217          * Get all magic words from cache.
3218          * @return array
3219          */
3220         function getMagicWords() {
3221                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3222         }
3223
3224         /**
3225          * Run the LanguageGetMagic hook once.
3226          */
3227         protected function doMagicHook() {
3228                 if ( $this->mMagicHookDone ) {
3229                         return;
3230                 }
3231                 $this->mMagicHookDone = true;
3232                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ], '1.16' );
3233         }
3234
3235         /**
3236          * Fill a MagicWord object with data from here
3237          *
3238          * @param MagicWord $mw
3239          */
3240         function getMagic( $mw ) {
3241                 // Saves a function call
3242                 if ( !$this->mMagicHookDone ) {
3243                         $this->doMagicHook();
3244                 }
3245
3246                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3247                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3248                 } else {
3249                         $rawEntry = self::$dataCache->getSubitem(
3250                                 $this->mCode, 'magicWords', $mw->mId );
3251                 }
3252
3253                 if ( !is_array( $rawEntry ) ) {
3254                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3255                 } else {
3256                         $mw->mCaseSensitive = $rawEntry[0];
3257                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3258                 }
3259         }
3260
3261         /**
3262          * Add magic words to the extension array
3263          *
3264          * @param array $newWords
3265          */
3266         function addMagicWordsByLang( $newWords ) {
3267                 $fallbackChain = $this->getFallbackLanguages();
3268                 $fallbackChain = array_reverse( $fallbackChain );
3269                 foreach ( $fallbackChain as $code ) {
3270                         if ( isset( $newWords[$code] ) ) {
3271                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3272                         }
3273                 }
3274         }
3275
3276         /**
3277          * Get special page names, as an associative array
3278          *   canonical name => array of valid names, including aliases
3279          * @return array
3280          */
3281         function getSpecialPageAliases() {
3282                 // Cache aliases because it may be slow to load them
3283                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3284                         // Initialise array
3285                         $this->mExtendedSpecialPageAliases =
3286                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3287                         Hooks::run( 'LanguageGetSpecialPageAliases',
3288                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ], '1.16' );
3289                 }
3290
3291                 return $this->mExtendedSpecialPageAliases;
3292         }
3293
3294         /**
3295          * Italic is unsuitable for some languages
3296          *
3297          * @param string $text The text to be emphasized.
3298          * @return string
3299          */
3300         function emphasize( $text ) {
3301                 return "<em>$text</em>";
3302         }
3303
3304         /**
3305          * Normally we output all numbers in plain en_US style, that is
3306          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3307          * point twohundredthirtyfive. However this is not suitable for all
3308          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3309          * Icelandic just want to use commas instead of dots, and dots instead
3310          * of commas like "293.291,235".
3311          *
3312          * An example of this function being called:
3313          * <code>
3314          * wfMessage( 'message' )->numParams( $num )->text()
3315          * </code>
3316          *
3317          * See $separatorTransformTable on MessageIs.php for
3318          * the , => . and . => , implementation.
3319          *
3320          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3321          * @param int|float $number The string to be formatted, should be an integer
3322          *   or a floating point number.
3323          * @param bool $nocommafy Set to true for special numbers like dates
3324          * @return string
3325          */
3326         public function formatNum( $number, $nocommafy = false ) {
3327                 global $wgTranslateNumerals;
3328                 if ( !$nocommafy ) {
3329                         $number = $this->commafy( $number );
3330                         $s = $this->separatorTransformTable();
3331                         if ( $s ) {
3332                                 $number = strtr( $number, $s );
3333                         }
3334                 }
3335
3336                 if ( $wgTranslateNumerals ) {
3337                         $s = $this->digitTransformTable();
3338                         if ( $s ) {
3339                                 $number = strtr( $number, $s );
3340                         }
3341                 }
3342
3343                 return (string)$number;
3344         }
3345
3346         /**
3347          * Front-end for non-commafied formatNum
3348          *
3349          * @param int|float $number The string to be formatted, should be an integer
3350          *        or a floating point number.
3351          * @since 1.21
3352          * @return string
3353          */
3354         public function formatNumNoSeparators( $number ) {
3355                 return $this->formatNum( $number, true );
3356         }
3357
3358         /**
3359          * @param string $number
3360          * @return string
3361          */
3362         public function parseFormattedNumber( $number ) {
3363                 $s = $this->digitTransformTable();
3364                 if ( $s ) {
3365                         // eliminate empty array values such as ''. (T66347)
3366                         $s = array_filter( $s );
3367                         $number = strtr( $number, array_flip( $s ) );
3368                 }
3369
3370                 $s = $this->separatorTransformTable();
3371                 if ( $s ) {
3372                         // eliminate empty array values such as ''. (T66347)
3373                         $s = array_filter( $s );
3374                         $number = strtr( $number, array_flip( $s ) );
3375                 }
3376
3377                 $number = strtr( $number, [ ',' => '' ] );
3378                 return $number;
3379         }
3380
3381         /**
3382          * Adds commas to a given number
3383          * @since 1.19
3384          * @param mixed $number
3385          * @return string
3386          */
3387         function commafy( $number ) {
3388                 $digitGroupingPattern = $this->digitGroupingPattern();
3389                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3390                 if ( $number === null ) {
3391                         return '';
3392                 }
3393
3394                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3395                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3396                         // In some languages it's conventional not to insert a thousands separator
3397                         // in numbers that are four digits long (1000-9999).
3398                         if ( $minimumGroupingDigits ) {
3399                                 // Number of '#' characters after last comma in the grouping pattern.
3400                                 // The pattern is hardcoded here, but this would vary for different patterns.
3401                                 $primaryGroupingSize = 3;
3402                                 // Maximum length of a number to suppress digit grouping for.
3403                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3404                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3405                                         return $number;
3406                                 }
3407                         }
3408                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3409                 } else {
3410                         // Ref: http://cldr.unicode.org/translation/number-patterns
3411                         $sign = "";
3412                         if ( intval( $number ) < 0 ) {
3413                                 // For negative numbers apply the algorithm like positive number and add sign.
3414                                 $sign = "-";
3415                                 $number = substr( $number, 1 );
3416                         }
3417                         $integerPart = [];
3418                         $decimalPart = [];
3419                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3420                         preg_match( "/\d+/", $number, $integerPart );
3421                         preg_match( "/\.\d*/", $number, $decimalPart );
3422                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3423                         if ( $groupedNumber === $number ) {
3424                                 // the string does not have any number part. Eg: .12345
3425                                 return $sign . $groupedNumber;
3426                         }
3427                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3428                         while ( $start > 0 ) {
3429                                 $match = $matches[0][$numMatches - 1];
3430                                 $matchLen = strlen( $match );
3431                                 $start = $end - $matchLen;
3432                                 if ( $start < 0 ) {
3433                                         $start = 0;
3434                                 }
3435                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3436                                 $end = $start;
3437                                 if ( $numMatches > 1 ) {
3438                                         // use the last pattern for the rest of the number
3439                                         $numMatches--;
3440                                 }
3441                                 if ( $start > 0 ) {
3442                                         $groupedNumber = "," . $groupedNumber;
3443                                 }
3444                         }
3445                         return $sign . $groupedNumber;
3446                 }
3447         }
3448
3449         /**
3450          * @return string
3451          */
3452         function digitGroupingPattern() {
3453                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3454         }
3455
3456         /**
3457          * @return array
3458          */
3459         function digitTransformTable() {
3460                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3461         }
3462
3463         /**
3464          * @return array
3465          */
3466         function separatorTransformTable() {
3467                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3468         }
3469
3470         /**
3471          * @return int|null
3472          */
3473         function minimumGroupingDigits() {
3474                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3475         }
3476
3477         /**
3478          * Take a list of strings and build a locale-friendly comma-separated
3479          * list, using the local comma-separator message.
3480          * The last two strings are chained with an "and".
3481          *
3482          * @param string[] $list
3483          * @return string
3484          */
3485         public function listToText( array $list ) {
3486                 $itemCount = count( $list );
3487                 if ( $itemCount < 1 ) {
3488                         return '';
3489                 }
3490                 $text = array_pop( $list );
3491                 if ( $itemCount > 1 ) {
3492                         $and = $this->msg( 'and' )->escaped();
3493                         $space = $this->msg( 'word-separator' )->escaped();
3494                         $comma = '';
3495                         if ( $itemCount > 2 ) {
3496                                 $comma = $this->msg( 'comma-separator' )->escaped();
3497                         }
3498                         $text = implode( $comma, $list ) . $and . $space . $text;
3499                 }
3500                 return $text;
3501         }
3502
3503         /**
3504          * Take a list of strings and build a locale-friendly comma-separated
3505          * list, using the local comma-separator message.
3506          * @param string[] $list Array of strings to put in a comma list
3507          * @return string
3508          */
3509         function commaList( array $list ) {
3510                 return implode(
3511                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3512                         $list
3513                 );
3514         }
3515
3516         /**
3517          * Take a list of strings and build a locale-friendly semicolon-separated
3518          * list, using the local semicolon-separator message.
3519          * @param string[] $list Array of strings to put in a semicolon list
3520          * @return string
3521          */
3522         function semicolonList( array $list ) {
3523                 return implode(
3524                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3525                         $list
3526                 );
3527         }
3528
3529         /**
3530          * Same as commaList, but separate it with the pipe instead.
3531          * @param string[] $list Array of strings to put in a pipe list
3532          * @return string
3533          */
3534         function pipeList( array $list ) {
3535                 return implode(
3536                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3537                         $list
3538                 );
3539         }
3540
3541         /**
3542          * This method is deprecated since 1.31 and kept as alias for truncateForDatabase, which
3543          * has replaced it. This method provides truncation suitable for DB.
3544          *
3545          * The database offers limited byte lengths for some columns in the database;
3546          * multi-byte character sets mean we need to ensure that only whole characters
3547          * are included, otherwise broken characters can be passed to the user.
3548          *
3549          * @deprecated since 1.31, use truncateForDatabase or truncateForVisual as appropriate.
3550          *
3551          * @param string $string String to truncate
3552          * @param int $length Maximum length (including ellipsis)
3553          * @param string $ellipsis String to append to the truncated text
3554          * @param bool $adjustLength Subtract length of ellipsis from $length.
3555          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3556          * @return string
3557          */
3558         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3559                 wfDeprecated( __METHOD__, '1.31' );
3560                 return $this->truncateForDatabase( $string, $length, $ellipsis, $adjustLength );
3561         }
3562
3563         /**
3564          * Truncate a string to a specified length in bytes, appending an optional
3565          * string (e.g. for ellipsis)
3566          *
3567          * If $length is negative, the string will be truncated from the beginning
3568          *
3569          * @since 1.31
3570          *
3571          * @param string $string String to truncate
3572          * @param int $length Maximum length in bytes
3573          * @param string $ellipsis String to append to the end of truncated text
3574          * @param bool $adjustLength Subtract length of ellipsis from $length
3575          *
3576          * @return string
3577          */
3578         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3579                 return $this->truncateInternal(
3580                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3581                 );
3582         }
3583
3584         /**
3585          * Truncate a string to a specified number of characters, appending an optional
3586          * string (e.g. for ellipsis).
3587          *
3588          * This provides multibyte version of truncate() method of this class, suitable for truncation
3589          * based on number of characters, instead of number of bytes.
3590          *
3591          * If $length is negative, the string will be truncated from the beginning.
3592          *
3593          * @since 1.31
3594          *
3595          * @param string $string String to truncate
3596          * @param int $length Maximum number of characters
3597          * @param string $ellipsis String to append to the end of truncated text
3598          * @param bool $adjustLength Subtract length of ellipsis from $length
3599          *
3600          * @return string
3601          */
3602         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3603                 // Passing encoding to mb_strlen and mb_substr is optional.
3604                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3605                 // explicit specification of encoding is skipped.
3606                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3607                 return $this->truncateInternal(
3608                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3609                 );
3610         }
3611
3612         /**
3613          * Internal method used for truncation. This method abstracts text truncation into
3614          * one common method, allowing users to provide length measurement function and
3615          * function for finding substring.
3616          *
3617          * For usages, see truncateForDatabase and truncateForVisual.
3618          *
3619          * @param string $string String to truncate
3620          * @param int $length Maximum length of final text
3621          * @param string $ellipsis String to append to the end of truncated text
3622          * @param bool $adjustLength Subtract length of ellipsis from $length
3623          * @param callable $measureLength Callable function used for determining the length of text
3624          * @param callable $getSubstring Callable function used for getting the substrings
3625          *
3626          * @return string
3627          */
3628         private function truncateInternal(
3629                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3630         ) {
3631                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3632                         throw new InvalidArgumentException( 'Invalid callback provided' );
3633                 }
3634
3635                 # Check if there is no need to truncate
3636                 if ( $measureLength( $string ) <= abs( $length ) ) {
3637                         return $string; // no need to truncate
3638                 }
3639
3640                 # Use the localized ellipsis character
3641                 if ( $ellipsis == '...' ) {
3642                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3643                 }
3644                 if ( $length == 0 ) {
3645                         return $ellipsis; // convention
3646                 }
3647
3648                 $stringOriginal = $string;
3649                 # If ellipsis length is >= $length then we can't apply $adjustLength
3650                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3651                         $string = $ellipsis; // this can be slightly unexpected
3652                 # Otherwise, truncate and add ellipsis...
3653                 } else {
3654                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3655                         if ( $length > 0 ) {
3656                                 $length -= $ellipsisLength;
3657                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3658                                 $string = $this->removeBadCharLast( $string );
3659                                 $string = rtrim( $string );
3660                                 $string = $string . $ellipsis;
3661                         } else {
3662                                 $length += $ellipsisLength;
3663                                 $string = $getSubstring( $string, $length ); // ...xyz
3664                                 $string = $this->removeBadCharFirst( $string );
3665                                 $string = ltrim( $string );
3666                                 $string = $ellipsis . $string;
3667                         }
3668                 }
3669
3670                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3671                 # This check is *not* redundant if $adjustLength, due to the single case where
3672                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3673                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3674                         return $string;
3675                 } else {
3676                         return $stringOriginal;
3677                 }
3678         }
3679
3680         /**
3681          * Remove bytes that represent an incomplete Unicode character
3682          * at the end of string (e.g. bytes of the char are missing)
3683          *
3684          * @param string $string
3685          * @return string
3686          */
3687         protected function removeBadCharLast( $string ) {
3688                 if ( $string != '' ) {
3689                         $char = ord( $string[strlen( $string ) - 1] );
3690                         $m = [];
3691                         if ( $char >= 0xc0 ) {
3692                                 # We got the first byte only of a multibyte char; remove it.
3693                                 $string = substr( $string, 0, -1 );
3694                         } elseif ( $char >= 0x80 &&
3695                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3696                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3697                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3698                         ) {
3699                                 # We chopped in the middle of a character; remove it
3700                                 $string = $m[1];
3701                         }
3702                 }
3703                 return $string;
3704         }
3705
3706         /**
3707          * Remove bytes that represent an incomplete Unicode character
3708          * at the start of string (e.g. bytes of the char are missing)
3709          *
3710          * @param string $string
3711          * @return string
3712          */
3713         protected function removeBadCharFirst( $string ) {
3714                 if ( $string != '' ) {
3715                         $char = ord( $string[0] );
3716                         if ( $char >= 0x80 && $char < 0xc0 ) {
3717                                 # We chopped in the middle of a character; remove the whole thing
3718                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3719                         }
3720                 }
3721                 return $string;
3722         }
3723
3724         /**
3725          * Truncate a string of valid HTML to a specified length in bytes,
3726          * appending an optional string (e.g. for ellipses), and return valid HTML
3727          *
3728          * This is only intended for styled/linked text, such as HTML with
3729          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3730          * Also, this will not detect things like "display:none" CSS.
3731          *
3732          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3733          *
3734          * @param string $text HTML string to truncate
3735          * @param int $length (zero/positive) Maximum length (including ellipses)
3736          * @param string $ellipsis String to append to the truncated text
3737          * @return string
3738          */
3739         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3740                 # Use the localized ellipsis character
3741                 if ( $ellipsis == '...' ) {
3742                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3743                 }
3744                 # Check if there is clearly no need to truncate
3745                 if ( $length <= 0 ) {
3746                         return $ellipsis; // no text shown, nothing to format (convention)
3747                 } elseif ( strlen( $text ) <= $length ) {
3748                         return $text; // string short enough even *with* HTML (short-circuit)
3749                 }
3750
3751                 $dispLen = 0; // innerHTML legth so far
3752                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3753                 $tagType = 0; // 0-open, 1-close
3754                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3755                 $entityState = 0; // 0-not entity, 1-entity
3756                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3757                 $openTags = []; // open tag stack
3758                 $maybeState = null; // possible truncation state
3759
3760                 $textLen = strlen( $text );
3761                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3762                 for ( $pos = 0; true; ++$pos ) {
3763                         # Consider truncation once the display length has reached the maximim.
3764                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3765                         # Check that we're not in the middle of a bracket/entity...
3766                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3767                                 if ( !$testingEllipsis ) {
3768                                         $testingEllipsis = true;
3769                                         # Save where we are; we will truncate here unless there turn out to
3770                                         # be so few remaining characters that truncation is not necessary.
3771                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3772                                                 $maybeState = [ $ret, $openTags ]; // save state
3773                                         }
3774                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3775                                         # String in fact does need truncation, the truncation point was OK.
3776                                         list( $ret, $openTags ) = $maybeState; // reload state
3777                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3778                                         $ret .= $ellipsis; // add ellipsis
3779                                         break;
3780                                 }
3781                         }
3782                         if ( $pos >= $textLen ) {
3783                                 break; // extra iteration just for above checks
3784                         }
3785
3786                         # Read the next char...
3787                         $ch = $text[$pos];
3788                         $lastCh = $pos ? $text[$pos - 1] : '';
3789                         $ret .= $ch; // add to result string
3790                         if ( $ch == '<' ) {
3791                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3792                                 $entityState = 0; // for bad HTML
3793                                 $bracketState = 1; // tag started (checking for backslash)
3794                         } elseif ( $ch == '>' ) {
3795                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3796                                 $entityState = 0; // for bad HTML
3797                                 $bracketState = 0; // out of brackets
3798                         } elseif ( $bracketState == 1 ) {
3799                                 if ( $ch == '/' ) {
3800                                         $tagType = 1; // close tag (e.g. "</span>")
3801                                 } else {
3802                                         $tagType = 0; // open tag (e.g. "<span>")
3803                                         $tag .= $ch;
3804                                 }
3805                                 $bracketState = 2; // building tag name
3806                         } elseif ( $bracketState == 2 ) {
3807                                 if ( $ch != ' ' ) {
3808                                         $tag .= $ch;
3809                                 } else {
3810                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3811                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3812                                 }
3813                         } elseif ( $bracketState == 0 ) {
3814                                 if ( $entityState ) {
3815                                         if ( $ch == ';' ) {
3816                                                 $entityState = 0;
3817                                                 $dispLen++; // entity is one displayed char
3818                                         }
3819                                 } else {
3820                                         if ( $neLength == 0 && !$maybeState ) {
3821                                                 // Save state without $ch. We want to *hit* the first
3822                                                 // display char (to get tags) but not *use* it if truncating.
3823                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3824                                         }
3825                                         if ( $ch == '&' ) {
3826                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3827                                         } else {
3828                                                 $dispLen++; // this char is displayed
3829                                                 // Add the next $max display text chars after this in one swoop...
3830                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3831                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3832                                                 $dispLen += $skipped;
3833                                                 $pos += $skipped;
3834                                         }
3835                                 }
3836                         }
3837                 }
3838                 // Close the last tag if left unclosed by bad HTML
3839                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3840                 while ( count( $openTags ) > 0 ) {
3841                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3842                 }
3843                 return $ret;
3844         }
3845
3846         /**
3847          * truncateHtml() helper function
3848          * like strcspn() but adds the skipped chars to $ret
3849          *
3850          * @param string $ret
3851          * @param string $text
3852          * @param string $search
3853          * @param int $start
3854          * @param null|int $len
3855          * @return int
3856          */
3857         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3858                 if ( $len === null ) {
3859                         $len = -1; // -1 means "no limit" for strcspn
3860                 } elseif ( $len < 0 ) {
3861                         $len = 0; // sanity
3862                 }
3863                 $skipCount = 0;
3864                 if ( $start < strlen( $text ) ) {
3865                         $skipCount = strcspn( $text, $search, $start, $len );
3866                         $ret .= substr( $text, $start, $skipCount );
3867                 }
3868                 return $skipCount;
3869         }
3870
3871         /**
3872          * truncateHtml() helper function
3873          * (a) push or pop $tag from $openTags as needed
3874          * (b) clear $tag value
3875          * @param string &$tag Current HTML tag name we are looking at
3876          * @param int $tagType (0-open tag, 1-close tag)
3877          * @param string $lastCh Character before the '>' that ended this tag
3878          * @param array &$openTags Open tag stack (not accounting for $tag)
3879          */
3880         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3881                 $tag = ltrim( $tag );
3882                 if ( $tag != '' ) {
3883                         if ( $tagType == 0 && $lastCh != '/' ) {
3884                                 $openTags[] = $tag; // tag opened (didn't close itself)
3885                         } elseif ( $tagType == 1 ) {
3886                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3887                                         array_pop( $openTags ); // tag closed
3888                                 }
3889                         }
3890                         $tag = '';
3891                 }
3892         }
3893
3894         /**
3895          * Grammatical transformations, needed for inflected languages
3896          * Invoked by putting {{grammar:case|word}} in a message
3897          *
3898          * @param string $word
3899          * @param string $case
3900          * @return string
3901          */
3902         function convertGrammar( $word, $case ) {
3903                 global $wgGrammarForms;
3904                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3905                         return $wgGrammarForms[$this->getCode()][$case][$word];
3906                 }
3907
3908                 $grammarTransformations = $this->getGrammarTransformations();
3909
3910                 if ( isset( $grammarTransformations[$case] ) ) {
3911                         $forms = $grammarTransformations[$case];
3912
3913                         // Some names of grammar rules are aliases for other rules.
3914                         // In such cases the value is a string rather than object,
3915                         // so load the actual rules.
3916                         if ( is_string( $forms ) ) {
3917                                 $forms = $grammarTransformations[$forms];
3918                         }
3919
3920                         foreach ( array_values( $forms ) as $rule ) {
3921                                 $form = $rule[0];
3922
3923                                 if ( $form === '@metadata' ) {
3924                                         continue;
3925                                 }
3926
3927                                 $replacement = $rule[1];
3928
3929                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3930                                 $patternMatches = preg_match( $regex, $word );
3931
3932                                 if ( $patternMatches === false ) {
3933                                         wfLogWarning(
3934                                                 'An error occurred while processing grammar. ' .
3935                                                 "Word: '$word'. Regex: /$form/."
3936                                         );
3937                                 } elseif ( $patternMatches === 1 ) {
3938                                         $word = preg_replace( $regex, $replacement, $word );
3939
3940                                         break;
3941                                 }
3942                         }
3943                 }
3944
3945                 return $word;
3946         }
3947
3948         /**
3949          * Get the grammar forms for the content language
3950          * @return array Array of grammar forms
3951          * @since 1.20
3952          */
3953         function getGrammarForms() {
3954                 global $wgGrammarForms;
3955                 if ( isset( $wgGrammarForms[$this->getCode()] )
3956                         && is_array( $wgGrammarForms[$this->getCode()] )
3957                 ) {
3958                         return $wgGrammarForms[$this->getCode()];
3959                 }
3960
3961                 return [];
3962         }
3963
3964         /**
3965          * Get the grammar transformations data for the language.
3966          * Used like grammar forms, with {{GRAMMAR}} and cases,
3967          * but uses pairs of regexes and replacements instead of code.
3968          *
3969          * @return array[] Array of grammar transformations.
3970          * @throws MWException
3971          * @since 1.28
3972          */
3973         public function getGrammarTransformations() {
3974                 $languageCode = $this->getCode();
3975
3976                 if ( self::$grammarTransformations === null ) {
3977                         self::$grammarTransformations = new MapCacheLRU( 10 );
3978                 }
3979
3980                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3981                         return self::$grammarTransformations->get( $languageCode );
3982                 }
3983
3984                 $data = [];
3985
3986                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3987                 if ( is_readable( $grammarDataFile ) ) {
3988                         $data = FormatJson::decode(
3989                                 file_get_contents( $grammarDataFile ),
3990                                 true
3991                         );
3992
3993                         if ( $data === null ) {
3994                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3995                         }
3996
3997                         self::$grammarTransformations->set( $languageCode, $data );
3998                 }
3999
4000                 return $data;
4001         }
4002
4003         /**
4004          * Provides an alternative text depending on specified gender.
4005          * Usage {{gender:username|masculine|feminine|unknown}}.
4006          * username is optional, in which case the gender of current user is used,
4007          * but only in (some) interface messages; otherwise default gender is used.
4008          *
4009          * If no forms are given, an empty string is returned. If only one form is
4010          * given, it will be returned unconditionally. These details are implied by
4011          * the caller and cannot be overridden in subclasses.
4012          *
4013          * If three forms are given, the default is to use the third (unknown) form.
4014          * If fewer than three forms are given, the default is to use the first (masculine) form.
4015          * These details can be overridden in subclasses.
4016          *
4017          * @param string $gender
4018          * @param array $forms
4019          *
4020          * @return string
4021          */
4022         function gender( $gender, $forms ) {
4023                 if ( !count( $forms ) ) {
4024                         return '';
4025                 }
4026                 $forms = $this->preConvertPlural( $forms, 2 );
4027                 if ( $gender === 'male' ) {
4028                         return $forms[0];
4029                 }
4030                 if ( $gender === 'female' ) {
4031                         return $forms[1];
4032                 }
4033                 return $forms[2] ?? $forms[0];
4034         }
4035
4036         /**
4037          * Plural form transformations, needed for some languages.
4038          * For example, there are 3 form of plural in Russian and Polish,
4039          * depending on "count mod 10". See [[w:Plural]]
4040          * For English it is pretty simple.
4041          *
4042          * Invoked by putting {{plural:count|wordform1|wordform2}}
4043          * or {{plural:count|wordform1|wordform2|wordform3}}
4044          *
4045          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
4046          *
4047          * @param int $count Non-localized number
4048          * @param array $forms Different plural forms
4049          * @return string Correct form of plural for $count in this language
4050          */
4051         function convertPlural( $count, $forms ) {
4052                 // Handle explicit n=pluralform cases
4053                 $forms = $this->handleExplicitPluralForms( $count, $forms );
4054                 if ( is_string( $forms ) ) {
4055                         return $forms;
4056                 }
4057                 if ( !count( $forms ) ) {
4058                         return '';
4059                 }
4060
4061                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
4062                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
4063                 return $forms[$pluralForm];
4064         }
4065
4066         /**
4067          * Handles explicit plural forms for Language::convertPlural()
4068          *
4069          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4070          * If an explicitly defined plural form matches the $count, then
4071          * string value returned, otherwise array returned for further consideration
4072          * by CLDR rules or overridden convertPlural().
4073          *
4074          * @since 1.23
4075          *
4076          * @param int $count Non-localized number
4077          * @param array $forms Different plural forms
4078          *
4079          * @return array|string
4080          */
4081         protected function handleExplicitPluralForms( $count, array $forms ) {
4082                 foreach ( $forms as $index => $form ) {
4083                         if ( preg_match( '/\d+=/i', $form ) ) {
4084                                 $pos = strpos( $form, '=' );
4085                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4086                                         return substr( $form, $pos + 1 );
4087                                 }
4088                                 unset( $forms[$index] );
4089                         }
4090                 }
4091                 return array_values( $forms );
4092         }
4093
4094         /**
4095          * Checks that convertPlural was given an array and pads it to requested
4096          * amount of forms by copying the last one.
4097          *
4098          * @param array $forms Array of forms given to convertPlural
4099          * @param int $count How many forms should there be at least
4100          * @return array Padded array of forms or an exception if not an array
4101          */
4102         protected function preConvertPlural( /* Array */ $forms, $count ) {
4103                 while ( count( $forms ) < $count ) {
4104                         $forms[] = $forms[count( $forms ) - 1];
4105                 }
4106                 return $forms;
4107         }
4108
4109         /**
4110          * Wraps argument with unicode control characters for directionality safety
4111          *
4112          * This solves the problem where directionality-neutral characters at the edge of
4113          * the argument string get interpreted with the wrong directionality from the
4114          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4115          *
4116          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4117          * directionality of the argument string, using the BIDI algorithm's own "First
4118          * strong directional codepoint" rule. Essentially, this works round the fact that
4119          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4120          * direction inference). The latter is cleaner but still not widely supported.
4121          *
4122          * @param string $text Text to wrap
4123          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4124          */
4125         public function embedBidi( $text = '' ) {
4126                 $dir = self::strongDirFromContent( $text );
4127                 if ( $dir === 'ltr' ) {
4128                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4129                         return self::$lre . $text . self::$pdf;
4130                 }
4131                 if ( $dir === 'rtl' ) {
4132                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4133                         return self::$rle . $text . self::$pdf;
4134                 }
4135                 // No strong directionality: do not wrap
4136                 return $text;
4137         }
4138
4139         /**
4140          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4141          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4142          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4143          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4144          * match up with it.
4145          *
4146          * @param string $str The validated block duration in English
4147          * @param User|null $user User object to use timezone from or null for $wgUser
4148          * @param int $now Current timestamp, for formatting relative block durations
4149          * @return string Somehow translated block duration
4150          * @see LanguageFi.php for example implementation
4151          */
4152         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4153                 $duration = SpecialBlock::getSuggestedDurations( $this );
4154                 foreach ( $duration as $show => $value ) {
4155                         if ( strcmp( $str, $value ) == 0 ) {
4156                                 return htmlspecialchars( trim( $show ) );
4157                         }
4158                 }
4159
4160                 if ( wfIsInfinity( $str ) ) {
4161                         foreach ( $duration as $show => $value ) {
4162                                 if ( wfIsInfinity( $value ) ) {
4163                                         return htmlspecialchars( trim( $show ) );
4164                                 }
4165                         }
4166                 }
4167
4168                 // If all else fails, return a standard duration or timestamp description.
4169                 $time = strtotime( $str, $now );
4170                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4171                         return $str;
4172                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4173                         // The result differs based on current time, so the difference
4174                         // is a fixed duration length.
4175                         return $this->formatDuration( $time - $now );
4176                 } else { // It's an absolute timestamp.
4177                         if ( $time === 0 ) {
4178                                 // wfTimestamp() handles 0 as current time instead of epoch.
4179                                 $time = '19700101000000';
4180                         }
4181                         if ( $user ) {
4182                                 return $this->userTimeAndDate( $time, $user );
4183                         }
4184                         return $this->timeanddate( $time );
4185                 }
4186         }
4187
4188         /**
4189          * languages like Chinese need to be segmented in order for the diff
4190          * to be of any use
4191          *
4192          * @param string $text
4193          * @return string
4194          */
4195         public function segmentForDiff( $text ) {
4196                 return $text;
4197         }
4198
4199         /**
4200          * and unsegment to show the result
4201          *
4202          * @param string $text
4203          * @return string
4204          */
4205         public function unsegmentForDiff( $text ) {
4206                 return $text;
4207         }
4208
4209         /**
4210          * Return the LanguageConverter used in the Language
4211          *
4212          * @since 1.19
4213          * @return LanguageConverter
4214          */
4215         public function getConverter() {
4216                 return $this->mConverter;
4217         }
4218
4219         /**
4220          * convert text to a variant
4221          *
4222          * @param string $text text to convert
4223          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4224          *      variant (if logged in), or the project default variant
4225          * @return string the converted string
4226          */
4227         public function autoConvert( $text, $variant = false ) {
4228                 return $this->mConverter->autoConvert( $text, $variant );
4229         }
4230
4231         /**
4232          * convert text to all supported variants
4233          *
4234          * @param string $text
4235          * @return array
4236          */
4237         public function autoConvertToAllVariants( $text ) {
4238                 return $this->mConverter->autoConvertToAllVariants( $text );
4239         }
4240
4241         /**
4242          * convert text to different variants of a language.
4243          *
4244          * @warning Glossary state is maintained between calls. This means
4245          *  if you pass unescaped text to this method it can cause an XSS
4246          *  in later calls to this method, even if the later calls have properly
4247          *  escaped the input. Never feed this method user controlled text that
4248          *  is not properly escaped!
4249          * @param string $text Content that has been already escaped for use in HTML
4250          * @return string HTML
4251          */
4252         public function convert( $text ) {
4253                 return $this->mConverter->convert( $text );
4254         }
4255
4256         /**
4257          * Convert a Title object to a string in the preferred variant
4258          *
4259          * @param Title $title
4260          * @return string
4261          */
4262         public function convertTitle( $title ) {
4263                 return $this->mConverter->convertTitle( $title );
4264         }
4265
4266         /**
4267          * Convert a namespace index to a string in the preferred variant
4268          *
4269          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4270          * @param string|null $variant variant to convert to, or null to use the user's preferred
4271          *      variant (if logged in), or the project default variant
4272          * @return string a string representation of the namespace
4273          */
4274         public function convertNamespace( $ns, $variant = null ) {
4275                 return $this->mConverter->convertNamespace( $ns, $variant );
4276         }
4277
4278         /**
4279          * Check if this is a language with variants
4280          *
4281          * @return bool
4282          */
4283         public function hasVariants() {
4284                 return count( $this->getVariants() ) > 1;
4285         }
4286
4287         /**
4288          * Check if the language has the specific variant
4289          *
4290          * @since 1.19
4291          * @param string $variant
4292          * @return bool
4293          */
4294         public function hasVariant( $variant ) {
4295                 return (bool)$this->mConverter->validateVariant( $variant );
4296         }
4297
4298         /**
4299          * Perform output conversion on a string, and encode for safe HTML output.
4300          * @param string $text Text to be converted
4301          * @param bool $isTitle Whether this conversion is for the article title
4302          * @return string
4303          * @todo this should get integrated somewhere sane
4304          */
4305         public function convertHtml( $text, $isTitle = false ) {
4306                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4307         }
4308
4309         /**
4310          * @param string $key
4311          * @return string
4312          */
4313         public function convertCategoryKey( $key ) {
4314                 return $this->mConverter->convertCategoryKey( $key );
4315         }
4316
4317         /**
4318          * Get the list of variants supported by this language
4319          * see sample implementation in LanguageZh.php
4320          *
4321          * @return string[] An array of language codes
4322          */
4323         public function getVariants() {
4324                 return $this->mConverter->getVariants();
4325         }
4326
4327         /**
4328          * @return string
4329          */
4330         public function getPreferredVariant() {
4331                 return $this->mConverter->getPreferredVariant();
4332         }
4333
4334         /**
4335          * @return string
4336          */
4337         public function getDefaultVariant() {
4338                 return $this->mConverter->getDefaultVariant();
4339         }
4340
4341         /**
4342          * @return string
4343          */
4344         public function getURLVariant() {
4345                 return $this->mConverter->getURLVariant();
4346         }
4347
4348         /**
4349          * If a language supports multiple variants, it is
4350          * possible that non-existing link in one variant
4351          * actually exists in another variant. this function
4352          * tries to find it. See e.g. LanguageZh.php
4353          * The input parameters may be modified upon return
4354          *
4355          * @param string &$link The name of the link
4356          * @param Title &$nt The title object of the link
4357          * @param bool $ignoreOtherCond To disable other conditions when
4358          *   we need to transclude a template or update a category's link
4359          */
4360         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4361                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4362         }
4363
4364         /**
4365          * returns language specific options used by User::getPageRenderHash()
4366          * for example, the preferred language variant
4367          *
4368          * @return string
4369          */
4370         function getExtraHashOptions() {
4371                 return $this->mConverter->getExtraHashOptions();
4372         }
4373
4374         /**
4375          * For languages that support multiple variants, the title of an
4376          * article may be displayed differently in different variants. this
4377          * function returns the apporiate title defined in the body of the article.
4378          *
4379          * @return string
4380          */
4381         public function getParsedTitle() {
4382                 return $this->mConverter->getParsedTitle();
4383         }
4384
4385         /**
4386          * Refresh the cache of conversion tables when
4387          * MediaWiki:Conversiontable* is updated.
4388          *
4389          * @param Title $title The Title of the page being updated
4390          */
4391         public function updateConversionTable( Title $title ) {
4392                 $this->mConverter->updateConversionTable( $title );
4393         }
4394
4395         /**
4396          * Prepare external link text for conversion. When the text is
4397          * a URL, it shouldn't be converted, and it'll be wrapped in
4398          * the "raw" tag (-{R| }-) to prevent conversion.
4399          *
4400          * This function is called "markNoConversion" for historical
4401          * reasons *BUT DIFFERS SIGNIFICANTLY* from
4402          * LanguageConverter::markNoConversion(), with which it is easily
4403          * confused.
4404          *
4405          * @param string $text Text to be used for external link
4406          * @param bool $noParse Wrap it without confirming it's a real URL first
4407          * @return string The tagged text
4408          * @deprecated since 1.32, use LanguageConverter::markNoConversion()
4409          *  instead.
4410          */
4411         public function markNoConversion( $text, $noParse = false ) {
4412                 wfDeprecated( __METHOD__, '1.32' );
4413                 // Excluding protocal-relative URLs may avoid many false positives.
4414                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4415                         return $this->mConverter->markNoConversion( $text );
4416                 } else {
4417                         return $text;
4418                 }
4419         }
4420
4421         /**
4422          * A regular expression to match legal word-trailing characters
4423          * which should be merged onto a link of the form [[foo]]bar.
4424          *
4425          * @return string
4426          */
4427         public function linkTrail() {
4428                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4429         }
4430
4431         /**
4432          * A regular expression character set to match legal word-prefixing
4433          * characters which should be merged onto a link of the form foo[[bar]].
4434          *
4435          * @return string
4436          */
4437         public function linkPrefixCharset() {
4438                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4439         }
4440
4441         /**
4442          * Get the "parent" language which has a converter to convert a "compatible" language
4443          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4444          *
4445          * @return Language|null
4446          * @since 1.22
4447          */
4448         public function getParentLanguage() {
4449                 if ( $this->mParentLanguage !== false ) {
4450                         return $this->mParentLanguage;
4451                 }
4452
4453                 $code = explode( '-', $this->getCode() )[0];
4454                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4455                         $this->mParentLanguage = null;
4456                         return null;
4457                 }
4458                 $lang = self::factory( $code );
4459                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4460                         $this->mParentLanguage = null;
4461                         return null;
4462                 }
4463
4464                 $this->mParentLanguage = $lang;
4465                 return $lang;
4466         }
4467
4468         /**
4469          * Compare with an other language object
4470          *
4471          * @since 1.28
4472          * @param Language $lang
4473          * @return bool
4474          */
4475         public function equals( Language $lang ) {
4476                 return $lang === $this || $lang->getCode() === $this->mCode;
4477         }
4478
4479         /**
4480          * Get the internal language code for this language object
4481          *
4482          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4483          * htmlspecialchars() or similar
4484          *
4485          * @return string
4486          */
4487         public function getCode() {
4488                 return $this->mCode;
4489         }
4490
4491         /**
4492          * Get the code in BCP 47 format which we can use
4493          * inside of html lang="" tags.
4494          *
4495          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4496          * htmlspecialchars() or similar.
4497          *
4498          * @since 1.19
4499          * @return string
4500          */
4501         public function getHtmlCode() {
4502                 if ( is_null( $this->mHtmlCode ) ) {
4503                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4504                 }
4505                 return $this->mHtmlCode;
4506         }
4507
4508         /**
4509          * @param string $code
4510          * @deprecated since 1.32, use Language::factory to create a new object instead.
4511          */
4512         public function setCode( $code ) {
4513                 $this->mCode = $code;
4514                 // Ensure we don't leave incorrect cached data lying around
4515                 $this->mHtmlCode = null;
4516                 $this->mParentLanguage = false;
4517         }
4518
4519         /**
4520          * Get the language code from a file name. Inverse of getFileName()
4521          * @param string $filename $prefix . $languageCode . $suffix
4522          * @param string $prefix Prefix before the language code
4523          * @param string $suffix Suffix after the language code
4524          * @return string Language code, or false if $prefix or $suffix isn't found
4525          */
4526         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4527                 $m = null;
4528                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4529                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4530                 if ( !count( $m ) ) {
4531                         return false;
4532                 }
4533                 return str_replace( '_', '-', strtolower( $m[1] ) );
4534         }
4535
4536         /**
4537          * @param string $code
4538          * @param bool $fallback Whether we're going through language fallback chain
4539          * @return string Name of the language class
4540          */
4541         public static function classFromCode( $code, $fallback = true ) {
4542                 if ( $fallback && $code == 'en' ) {
4543                         return 'Language';
4544                 } else {
4545                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4546                 }
4547         }
4548
4549         /**
4550          * Get the name of a file for a certain language code
4551          * @param string $prefix Prepend this to the filename
4552          * @param string $code Language code
4553          * @param string $suffix Append this to the filename
4554          * @throws MWException
4555          * @return string $prefix . $mangledCode . $suffix
4556          */
4557         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4558                 if ( !self::isValidBuiltInCode( $code ) ) {
4559                         throw new MWException( "Invalid language code \"$code\"" );
4560                 }
4561
4562                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4563         }
4564
4565         /**
4566          * @param string $code
4567          * @return string
4568          */
4569         public static function getMessagesFileName( $code ) {
4570                 global $IP;
4571                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4572                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4573                 return $file;
4574         }
4575
4576         /**
4577          * @param string $code
4578          * @return string
4579          * @throws MWException
4580          * @since 1.23
4581          */
4582         public static function getJsonMessagesFileName( $code ) {
4583                 global $IP;
4584
4585                 if ( !self::isValidBuiltInCode( $code ) ) {
4586                         throw new MWException( "Invalid language code \"$code\"" );
4587                 }
4588
4589                 return "$IP/languages/i18n/$code.json";
4590         }
4591
4592         /**
4593          * Get the first fallback for a given language.
4594          *
4595          * @param string $code
4596          *
4597          * @return bool|string
4598          */
4599         public static function getFallbackFor( $code ) {
4600                 $fallbacks = self::getFallbacksFor( $code );
4601                 if ( $fallbacks ) {
4602                         return $fallbacks[0];
4603                 }
4604                 return false;
4605         }
4606
4607         /**
4608          * Get the ordered list of fallback languages.
4609          *
4610          * @since 1.19
4611          * @param string $code Language code
4612          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4613          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4614          * @throws MWException
4615          * @return array List of language codes
4616          */
4617         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4618                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4619                         return [];
4620                 }
4621                 switch ( $mode ) {
4622                         case self::MESSAGES_FALLBACKS:
4623                                 // For unknown languages, fallbackSequence returns an empty array,
4624                                 // hardcode fallback to 'en' in that case as English messages are
4625                                 // always defined.
4626                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4627                         case self::STRICT_FALLBACKS:
4628                                 // Use this mode when you don't want to fallback to English unless
4629                                 // explicitly defined, for example when you have language-variant icons
4630                                 // and an international language-independent fallback.
4631                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4632                         default:
4633                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4634                 }
4635         }
4636
4637         /**
4638          * Get the ordered list of fallback languages, ending with the fallback
4639          * language chain for the site language.
4640          *
4641          * @since 1.22
4642          * @param string $code Language code
4643          * @return array Array( fallbacks, site fallbacks )
4644          */
4645         public static function getFallbacksIncludingSiteLanguage( $code ) {
4646                 global $wgLanguageCode;
4647
4648                 // Usually, we will only store a tiny number of fallback chains, so we
4649                 // keep them in static memory.
4650                 $cacheKey = "{$code}-{$wgLanguageCode}";
4651
4652                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4653                         $fallbacks = self::getFallbacksFor( $code );
4654
4655                         // Append the site's fallback chain, including the site language itself
4656                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4657                         array_unshift( $siteFallbacks, $wgLanguageCode );
4658
4659                         // Eliminate any languages already included in the chain
4660                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4661
4662                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4663                 }
4664                 return self::$fallbackLanguageCache[$cacheKey];
4665         }
4666
4667         /**
4668          * Get all messages for a given language
4669          * WARNING: this may take a long time. If you just need all message *keys*
4670          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4671          *
4672          * @param string $code
4673          *
4674          * @return array
4675          */
4676         public static function getMessagesFor( $code ) {
4677                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4678         }
4679
4680         /**
4681          * Get a message for a given language
4682          *
4683          * @param string $key
4684          * @param string $code
4685          *
4686          * @return string
4687          */
4688         public static function getMessageFor( $key, $code ) {
4689                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4690         }
4691
4692         /**
4693          * Get all message keys for a given language. This is a faster alternative to
4694          * array_keys( Language::getMessagesFor( $code ) )
4695          *
4696          * @since 1.19
4697          * @param string $code Language code
4698          * @return array Array of message keys (strings)
4699          */
4700         public static function getMessageKeysFor( $code ) {
4701                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4702         }
4703
4704         /**
4705          * @param string $talk
4706          * @return mixed
4707          */
4708         function fixVariableInNamespace( $talk ) {
4709                 if ( strpos( $talk, '$1' ) === false ) {
4710                         return $talk;
4711                 }
4712
4713                 global $wgMetaNamespace;
4714                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4715
4716                 # Allow grammar transformations
4717                 # Allowing full message-style parsing would make simple requests
4718                 # such as action=raw much more expensive than they need to be.
4719                 # This will hopefully cover most cases.
4720                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4721                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4722                 return str_replace( ' ', '_', $talk );
4723         }
4724
4725         /**
4726          * @param string $m
4727          * @return string
4728          */
4729         function replaceGrammarInNamespace( $m ) {
4730                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4731         }
4732
4733         /**
4734          * Decode an expiry (block, protection, etc) which has come from the DB
4735          *
4736          * @param string $expiry Database expiry String
4737          * @param bool|int $format True to process using language functions, or TS_ constant
4738          *     to return the expiry in a given timestamp
4739          * @param string $infinity If $format is not true, use this string for infinite expiry
4740          * @return string
4741          * @since 1.18
4742          */
4743         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4744                 static $dbInfinity;
4745                 if ( $dbInfinity === null ) {
4746                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4747                 }
4748
4749                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4750                         return $format === true
4751                                 ? $this->getMessageFromDB( 'infiniteblock' )
4752                                 : $infinity;
4753                 } else {
4754                         return $format === true
4755                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4756                                 : wfTimestamp( $format, $expiry );
4757                 }
4758         }
4759
4760         /**
4761          * Formats a time given in seconds into a string representation of that time.
4762          *
4763          * @param int|float $seconds
4764          * @param array $format An optional argument that formats the returned string in different ways:
4765          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4766          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4767          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4768          *     and friends.
4769          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4770          *     or 'avoidminutes'.
4771          * @return string
4772          */
4773         function formatTimePeriod( $seconds, $format = [] ) {
4774                 if ( !is_array( $format ) ) {
4775                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4776                 }
4777                 if ( !isset( $format['avoid'] ) ) {
4778                         $format['avoid'] = false;
4779                 }
4780                 if ( !isset( $format['noabbrevs'] ) ) {
4781                         $format['noabbrevs'] = false;
4782                 }
4783                 $secondsMsg = wfMessage(
4784                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4785                 $minutesMsg = wfMessage(
4786                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4787                 $hoursMsg = wfMessage(
4788                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4789                 $daysMsg = wfMessage(
4790                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4791
4792                 if ( round( $seconds * 10 ) < 100 ) {
4793                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4794                         $s = $secondsMsg->params( $s )->text();
4795                 } elseif ( round( $seconds ) < 60 ) {
4796                         $s = $this->formatNum( round( $seconds ) );
4797                         $s = $secondsMsg->params( $s )->text();
4798                 } elseif ( round( $seconds ) < 3600 ) {
4799                         $minutes = floor( $seconds / 60 );
4800                         $secondsPart = round( fmod( $seconds, 60 ) );
4801                         if ( $secondsPart == 60 ) {
4802                                 $secondsPart = 0;
4803                                 $minutes++;
4804                         }
4805                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4806                         $s .= ' ';
4807                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4808                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4809                         $hours = floor( $seconds / 3600 );
4810                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4811                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4812                         if ( $secondsPart == 60 ) {
4813                                 $secondsPart = 0;
4814                                 $minutes++;
4815                         }
4816                         if ( $minutes == 60 ) {
4817                                 $minutes = 0;
4818                                 $hours++;
4819                         }
4820                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4821                         $s .= ' ';
4822                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4823                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4824                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4825                         }
4826                 } else {
4827                         $days = floor( $seconds / 86400 );
4828                         if ( $format['avoid'] === 'avoidminutes' ) {
4829                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4830                                 if ( $hours == 24 ) {
4831                                         $hours = 0;
4832                                         $days++;
4833                                 }
4834                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4835                                 $s .= ' ';
4836                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4837                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4838                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4839                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4840                                 if ( $minutes == 60 ) {
4841                                         $minutes = 0;
4842                                         $hours++;
4843                                 }
4844                                 if ( $hours == 24 ) {
4845                                         $hours = 0;
4846                                         $days++;
4847                                 }
4848                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4849                                 $s .= ' ';
4850                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4851                                 $s .= ' ';
4852                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4853                         } else {
4854                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4855                                 $s .= ' ';
4856                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4857                         }
4858                 }
4859                 return $s;
4860         }
4861
4862         /**
4863          * Format a bitrate for output, using an appropriate
4864          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4865          *   the magnitude in question.
4866          *
4867          * This use base 1000. For base 1024 use formatSize(), for another base
4868          * see formatComputingNumbers().
4869          *
4870          * @param int $bps
4871          * @return string
4872          */
4873         function formatBitrate( $bps ) {
4874                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4875         }
4876
4877         /**
4878          * @param int $size Size of the unit
4879          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4880          * @param string $messageKey Message key to be uesd
4881          * @return string
4882          */
4883         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4884                 if ( $size <= 0 ) {
4885                         return str_replace( '$1', $this->formatNum( $size ),
4886                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4887                         );
4888                 }
4889                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4890                 $index = 0;
4891
4892                 $maxIndex = count( $sizes ) - 1;
4893                 while ( $size >= $boundary && $index < $maxIndex ) {
4894                         $index++;
4895                         $size /= $boundary;
4896                 }
4897
4898                 // For small sizes no decimal places necessary
4899                 $round = 0;
4900                 if ( $index > 1 ) {
4901                         // For MB and bigger two decimal places are smarter
4902                         $round = 2;
4903                 }
4904                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4905
4906                 $size = round( $size, $round );
4907                 $text = $this->getMessageFromDB( $msg );
4908                 return str_replace( '$1', $this->formatNum( $size ), $text );
4909         }
4910
4911         /**
4912          * Format a size in bytes for output, using an appropriate
4913          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4914          *
4915          * This method use base 1024. For base 1000 use formatBitrate(), for
4916          * another base see formatComputingNumbers()
4917          *
4918          * @param int $size Size to format
4919          * @return string Plain text (not HTML)
4920          */
4921         function formatSize( $size ) {
4922                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4923         }
4924
4925         /**
4926          * Make a list item, used by various special pages
4927          *
4928          * @param string $page Page link
4929          * @param string $details HTML safe text between brackets
4930          * @param bool $oppositedm Add the direction mark opposite to your
4931          *   language, to display text properly
4932          * @return string HTML escaped
4933          */
4934         function specialList( $page, $details, $oppositedm = true ) {
4935                 if ( !$details ) {
4936                         return $page;
4937                 }
4938
4939                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4940                 return $page .
4941                         $dirmark .
4942                         $this->msg( 'word-separator' )->escaped() .
4943                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4944         }
4945
4946         /**
4947          * Generate (prev x| next x) (20|50|100...) type links for paging
4948          *
4949          * @param Title $title Title object to link
4950          * @param int $offset
4951          * @param int $limit
4952          * @param array $query Optional URL query parameter string
4953          * @param bool $atend Optional param for specified if this is the last page
4954          * @return string
4955          */
4956         public function viewPrevNext( Title $title, $offset, $limit,
4957                 array $query = [], $atend = false
4958         ) {
4959                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4960
4961                 # Make 'previous' link
4962                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4963                 if ( $offset > 0 ) {
4964                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4965                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4966                 } else {
4967                         $plink = htmlspecialchars( $prev );
4968                 }
4969
4970                 # Make 'next' link
4971                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4972                 if ( $atend ) {
4973                         $nlink = htmlspecialchars( $next );
4974                 } else {
4975                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4976                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4977                 }
4978
4979                 # Make links to set number of items per page
4980                 $numLinks = [];
4981                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4982                         $numLinks[] = $this->numLink( $title, $offset, $num,
4983                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4984                 }
4985
4986                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4987                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4988         }
4989
4990         /**
4991          * Helper function for viewPrevNext() that generates links
4992          *
4993          * @param Title $title Title object to link
4994          * @param int $offset
4995          * @param int $limit
4996          * @param array $query Extra query parameters
4997          * @param string $link Text to use for the link; will be escaped
4998          * @param string $tooltipMsg Name of the message to use as tooltip
4999          * @param string $class Value of the "class" attribute of the link
5000          * @return string HTML fragment
5001          */
5002         private function numLink( Title $title, $offset, $limit, array $query, $link,
5003                 $tooltipMsg, $class
5004         ) {
5005                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
5006                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
5007                         ->numParams( $limit )->text();
5008
5009                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
5010                         'title' => $tooltip, 'class' => $class ], $link );
5011         }
5012
5013         /**
5014          * Get the conversion rule title, if any.
5015          *
5016          * @return string
5017          */
5018         public function getConvRuleTitle() {
5019                 return $this->mConverter->getConvRuleTitle();
5020         }
5021
5022         /**
5023          * Get the compiled plural rules for the language
5024          * @since 1.20
5025          * @return array Associative array with plural form, and plural rule as key-value pairs
5026          */
5027         public function getCompiledPluralRules() {
5028                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
5029                 $fallbacks = self::getFallbacksFor( $this->mCode );
5030                 if ( !$pluralRules ) {
5031                         foreach ( $fallbacks as $fallbackCode ) {
5032                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
5033                                 if ( $pluralRules ) {
5034                                         break;
5035                                 }
5036                         }
5037                 }
5038                 return $pluralRules;
5039         }
5040
5041         /**
5042          * Get the plural rules for the language
5043          * @since 1.20
5044          * @return array Associative array with plural form number and plural rule as key-value pairs
5045          */
5046         public function getPluralRules() {
5047                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
5048                 $fallbacks = self::getFallbacksFor( $this->mCode );
5049                 if ( !$pluralRules ) {
5050                         foreach ( $fallbacks as $fallbackCode ) {
5051                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
5052                                 if ( $pluralRules ) {
5053                                         break;
5054                                 }
5055                         }
5056                 }
5057                 return $pluralRules;
5058         }
5059
5060         /**
5061          * Get the plural rule types for the language
5062          * @since 1.22
5063          * @return array Associative array with plural form number and plural rule type as key-value pairs
5064          */
5065         public function getPluralRuleTypes() {
5066                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
5067                 $fallbacks = self::getFallbacksFor( $this->mCode );
5068                 if ( !$pluralRuleTypes ) {
5069                         foreach ( $fallbacks as $fallbackCode ) {
5070                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
5071                                 if ( $pluralRuleTypes ) {
5072                                         break;
5073                                 }
5074                         }
5075                 }
5076                 return $pluralRuleTypes;
5077         }
5078
5079         /**
5080          * Find the index number of the plural rule appropriate for the given number
5081          * @param int $number
5082          * @return int The index number of the plural rule
5083          */
5084         public function getPluralRuleIndexNumber( $number ) {
5085                 $pluralRules = $this->getCompiledPluralRules();
5086                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
5087                 return $form;
5088         }
5089
5090         /**
5091          * Find the plural rule type appropriate for the given number
5092          * For example, if the language is set to Arabic, getPluralType(5) should
5093          * return 'few'.
5094          * @since 1.22
5095          * @param int $number
5096          * @return string The name of the plural rule type, e.g. one, two, few, many
5097          */
5098         public function getPluralRuleType( $number ) {
5099                 $index = $this->getPluralRuleIndexNumber( $number );
5100                 $pluralRuleTypes = $this->getPluralRuleTypes();
5101                 if ( isset( $pluralRuleTypes[$index] ) ) {
5102                         return $pluralRuleTypes[$index];
5103                 } else {
5104                         return 'other';
5105                 }
5106         }
5107 }