'CliInstaller' => __DIR__ . '/includes/installer/CliInstaller.php',
'CloneDatabase' => __DIR__ . '/includes/db/CloneDatabase.php',
'CodeContentHandler' => __DIR__ . '/includes/content/CodeContentHandler.php',
- 'Collation' => __DIR__ . '/includes/Collation.php',
- 'CollationCkb' => __DIR__ . '/includes/Collation.php',
- 'CollationEt' => __DIR__ . '/includes/Collation.php',
+ 'Collation' => __DIR__ . '/includes/collation/Collation.php',
+ 'CollationCkb' => __DIR__ . '/includes/collation/CollationCkb.php',
+ 'CollationEt' => __DIR__ . '/includes/collation/CollationEt.php',
'CommandLineInc' => __DIR__ . '/maintenance/commandLine.inc',
'CommandLineInstaller' => __DIR__ . '/maintenance/install.php',
'CompareParserCache' => __DIR__ . '/maintenance/compareParserCache.php',
'EventRelayer' => __DIR__ . '/includes/libs/eventrelayer/EventRelayer.php',
'EventRelayerGroup' => __DIR__ . '/includes/EventRelayerGroup.php',
'EventRelayerMCRD' => __DIR__ . '/includes/libs/eventrelayer/EventRelayerMCRD.php',
- 'EventRelayerNull' => __DIR__ . '/includes/libs/eventrelayer/EventRelayer.php',
+ 'EventRelayerNull' => __DIR__ . '/includes/libs/eventrelayer/EventRelayerNull.php',
'Exif' => __DIR__ . '/includes/media/Exif.php',
'ExifBitmapHandler' => __DIR__ . '/includes/media/ExifBitmap.php',
'ExplodeIterator' => __DIR__ . '/includes/libs/ExplodeIterator.php',
'IPSet' => __DIR__ . '/includes/compat/IPSetCompat.php',
'IPTC' => __DIR__ . '/includes/media/IPTC.php',
'IRCColourfulRCFeedFormatter' => __DIR__ . '/includes/rcfeed/IRCColourfulRCFeedFormatter.php',
- 'IcuCollation' => __DIR__ . '/includes/Collation.php',
- 'IdentityCollation' => __DIR__ . '/includes/Collation.php',
+ 'IcuCollation' => __DIR__ . '/includes/collation/IcuCollation.php',
+ 'IdentityCollation' => __DIR__ . '/includes/collation/IdentityCollation.php',
'ImageBuilder' => __DIR__ . '/maintenance/rebuildImages.php',
'ImageCleanup' => __DIR__ . '/maintenance/cleanupImages.php',
'ImageGallery' => __DIR__ . '/includes/gallery/TraditionalImageGallery.php',
'UploadStashNotLoggedInException' => __DIR__ . '/includes/upload/UploadStash.php',
'UploadStashWrongOwnerException' => __DIR__ . '/includes/upload/UploadStash.php',
'UploadStashZeroLengthFileException' => __DIR__ . '/includes/upload/UploadStash.php',
- 'UppercaseCollation' => __DIR__ . '/includes/Collation.php',
+ 'UppercaseCollation' => __DIR__ . '/includes/collation/UppercaseCollation.php',
'UsageException' => __DIR__ . '/includes/api/ApiMain.php',
'User' => __DIR__ . '/includes/user/User.php',
'UserArray' => __DIR__ . '/includes/user/UserArray.php',
+++ /dev/null
-<?php
-/**
- * Database row sorting.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write to the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
- * http://www.gnu.org/copyleft/gpl.html
- *
- * @file
- */
-
-abstract class Collation {
- private static $instance;
-
- /**
- * @return Collation
- */
- static function singleton() {
- if ( !self::$instance ) {
- global $wgCategoryCollation;
- self::$instance = self::factory( $wgCategoryCollation );
- }
- return self::$instance;
- }
-
- /**
- * @throws MWException
- * @param string $collationName
- * @return Collation
- */
- static function factory( $collationName ) {
- switch ( $collationName ) {
- case 'uppercase':
- return new UppercaseCollation;
- case 'identity':
- return new IdentityCollation;
- case 'uca-default':
- return new IcuCollation( 'root' );
- case 'xx-uca-ckb':
- return new CollationCkb;
- case 'xx-uca-et':
- return new CollationEt;
- default:
- $match = [];
- if ( preg_match( '/^uca-([a-z@=-]+)$/', $collationName, $match ) ) {
- return new IcuCollation( $match[1] );
- }
-
- # Provide a mechanism for extensions to hook in.
- $collationObject = null;
- Hooks::run( 'Collation::factory', [ $collationName, &$collationObject ] );
-
- if ( $collationObject instanceof Collation ) {
- return $collationObject;
- }
-
- // If all else fails...
- throw new MWException( __METHOD__ . ": unknown collation type \"$collationName\"" );
- }
- }
-
- /**
- * Given a string, convert it to a (hopefully short) key that can be used
- * for efficient sorting. A binary sort according to the sortkeys
- * corresponds to a logical sort of the corresponding strings. Current
- * code expects that a line feed character should sort before all others, but
- * has no other particular expectations (and that one can be changed if
- * necessary).
- *
- * @param string $string UTF-8 string
- * @return string Binary sortkey
- */
- abstract function getSortKey( $string );
-
- /**
- * Given a string, return the logical "first letter" to be used for
- * grouping on category pages and so on. This has to be coordinated
- * carefully with convertToSortkey(), or else the sorted list might jump
- * back and forth between the same "initial letters" or other pathological
- * behavior. For instance, if you just return the first character, but "a"
- * sorts the same as "A" based on getSortKey(), then you might get a
- * list like
- *
- * == A ==
- * * [[Aardvark]]
- *
- * == a ==
- * * [[antelope]]
- *
- * == A ==
- * * [[Ape]]
- *
- * etc., assuming for the sake of argument that $wgCapitalLinks is false.
- *
- * @param string $string UTF-8 string
- * @return string UTF-8 string corresponding to the first letter of input
- */
- abstract function getFirstLetter( $string );
-}
-
-class UppercaseCollation extends Collation {
- private $lang;
-
- function __construct() {
- // Get a language object so that we can use the generic UTF-8 uppercase
- // function there
- $this->lang = Language::factory( 'en' );
- }
-
- function getSortKey( $string ) {
- return $this->lang->uc( $string );
- }
-
- function getFirstLetter( $string ) {
- if ( $string[0] == "\0" ) {
- $string = substr( $string, 1 );
- }
- return $this->lang->ucfirst( $this->lang->firstChar( $string ) );
- }
-}
-
-/**
- * Collation class that's essentially a no-op.
- *
- * Does sorting based on binary value of the string.
- * Like how things were pre 1.17.
- */
-class IdentityCollation extends Collation {
-
- function getSortKey( $string ) {
- return $string;
- }
-
- function getFirstLetter( $string ) {
- global $wgContLang;
- // Copied from UppercaseCollation.
- // I'm kind of unclear on when this could happen...
- if ( $string[0] == "\0" ) {
- $string = substr( $string, 1 );
- }
- return $wgContLang->firstChar( $string );
- }
-}
-
-class IcuCollation extends Collation {
- const FIRST_LETTER_VERSION = 2;
-
- /** @var Collator */
- private $primaryCollator;
-
- /** @var Collator */
- private $mainCollator;
-
- /** @var string */
- private $locale;
-
- /** @var Language */
- protected $digitTransformLanguage;
-
- /** @var array */
- private $firstLetterData;
-
- /**
- * Unified CJK blocks.
- *
- * The same definition of a CJK block must be used for both Collation and
- * generateCollationData.php. These blocks are omitted from the first
- * letter data, as an optimisation measure and because the default UCA table
- * is pretty useless for sorting Chinese text anyway. Japanese and Korean
- * blocks are not included here, because they are smaller and more useful.
- */
- private static $cjkBlocks = [
- [ 0x2E80, 0x2EFF ], // CJK Radicals Supplement
- [ 0x2F00, 0x2FDF ], // Kangxi Radicals
- [ 0x2FF0, 0x2FFF ], // Ideographic Description Characters
- [ 0x3000, 0x303F ], // CJK Symbols and Punctuation
- [ 0x31C0, 0x31EF ], // CJK Strokes
- [ 0x3200, 0x32FF ], // Enclosed CJK Letters and Months
- [ 0x3300, 0x33FF ], // CJK Compatibility
- [ 0x3400, 0x4DBF ], // CJK Unified Ideographs Extension A
- [ 0x4E00, 0x9FFF ], // CJK Unified Ideographs
- [ 0xF900, 0xFAFF ], // CJK Compatibility Ideographs
- [ 0xFE30, 0xFE4F ], // CJK Compatibility Forms
- [ 0x20000, 0x2A6DF ], // CJK Unified Ideographs Extension B
- [ 0x2A700, 0x2B73F ], // CJK Unified Ideographs Extension C
- [ 0x2B740, 0x2B81F ], // CJK Unified Ideographs Extension D
- [ 0x2F800, 0x2FA1F ], // CJK Compatibility Ideographs Supplement
- ];
-
- /**
- * Additional characters (or character groups) to be considered separate
- * letters for given languages, or to be removed from the list of such
- * letters (denoted by keys starting with '-').
- *
- * These are additions to (or subtractions from) the data stored in the
- * first-letters-root.ser file (which among others includes full basic latin,
- * cyrillic and greek alphabets).
- *
- * "Separate letter" is a letter that would have a separate heading/section
- * for it in a dictionary or a phone book in this language. This data isn't
- * used for sorting (the ICU library handles that), only for deciding which
- * characters (or character groups) to use as headings.
- *
- * Initially generated based on the primary level of Unicode collation
- * tailorings available at http://developer.mimer.com/charts/tailorings.htm ,
- * later modified.
- *
- * Empty arrays are intended; this signifies that the data for the language is
- * available and that there are, in fact, no additional letters to consider.
- */
- private static $tailoringFirstLetters = [
- // Verified by native speakers
- 'be' => [ "Ё" ],
- 'be-tarask' => [ "Ё" ],
- 'cy' => [ "Ch", "Dd", "Ff", "Ng", "Ll", "Ph", "Rh", "Th" ],
- 'en' => [],
- 'fa' => [ "آ", "ء", "ه" ],
- 'fi' => [ "Å", "Ä", "Ö" ],
- 'fr' => [],
- 'hu' => [ "Cs", "Dz", "Dzs", "Gy", "Ly", "Ny", "Ö", "Sz", "Ty", "Ü", "Zs" ],
- 'is' => [ "Á", "Ð", "É", "Í", "Ó", "Ú", "Ý", "Þ", "Æ", "Ö", "Å" ],
- 'it' => [],
- 'lv' => [ "Č", "Ģ", "Ķ", "Ļ", "Ņ", "Š", "Ž" ],
- 'pl' => [ "Ą", "Ć", "Ę", "Ł", "Ń", "Ó", "Ś", "Ź", "Ż" ],
- 'pt' => [],
- 'ru' => [],
- 'sv' => [ "Å", "Ä", "Ö" ],
- 'sv@collation=standard' => [ "Å", "Ä", "Ö" ],
- 'uk' => [ "Ґ", "Ь" ],
- 'vi' => [ "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ],
- // Not verified, but likely correct
- 'af' => [],
- 'ast' => [ "Ch", "Ll", "Ñ" ],
- 'az' => [ "Ç", "Ə", "Ğ", "İ", "Ö", "Ş", "Ü" ],
- 'bg' => [],
- 'br' => [ "Ch", "C'h" ],
- 'bs' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ],
- 'ca' => [],
- 'co' => [],
- 'cs' => [ "Č", "Ch", "Ř", "Š", "Ž" ],
- 'da' => [ "Æ", "Ø", "Å" ],
- 'de' => [],
- 'dsb' => [ "Č", "Ć", "Dź", "Ě", "Ch", "Ł", "Ń", "Ŕ", "Š", "Ś", "Ž", "Ź" ],
- 'el' => [],
- 'eo' => [ "Ĉ", "Ĝ", "Ĥ", "Ĵ", "Ŝ", "Ŭ" ],
- 'es' => [ "Ñ" ],
- 'et' => [ "Š", "Ž", "Õ", "Ä", "Ö", "Ü", "W" ], // added W for CollationEt (xx-uca-et)
- 'eu' => [ "Ñ" ],
- 'fo' => [ "Á", "Ð", "Í", "Ó", "Ú", "Ý", "Æ", "Ø", "Å" ],
- 'fur' => [ "À", "Á", "Â", "È", "Ì", "Ò", "Ù" ],
- 'fy' => [],
- 'ga' => [],
- 'gd' => [],
- 'gl' => [ "Ch", "Ll", "Ñ" ],
- 'hr' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ],
- 'hsb' => [ "Č", "Dź", "Ě", "Ch", "Ł", "Ń", "Ř", "Š", "Ć", "Ž" ],
- 'kk' => [ "Ү", "І" ],
- 'kl' => [ "Æ", "Ø", "Å" ],
- 'ku' => [ "Ç", "Ê", "Î", "Ş", "Û" ],
- 'ky' => [ "Ё" ],
- 'la' => [],
- 'lb' => [],
- 'lt' => [ "Č", "Š", "Ž" ],
- 'mk' => [],
- 'mo' => [ "Ă", "Â", "Î", "Ş", "Ţ" ],
- 'mt' => [ "Ċ", "Ġ", "Għ", "Ħ", "Ż" ],
- 'nl' => [],
- 'no' => [ "Æ", "Ø", "Å" ],
- 'oc' => [],
- 'rm' => [],
- 'ro' => [ "Ă", "Â", "Î", "Ş", "Ţ" ],
- 'rup' => [ "Ă", "Â", "Î", "Ľ", "Ń", "Ş", "Ţ" ],
- 'sco' => [],
- 'sk' => [ "Ä", "Č", "Ch", "Ô", "Š", "Ž" ],
- 'sl' => [ "Č", "Š", "Ž" ],
- 'smn' => [ "Á", "Č", "Đ", "Ŋ", "Š", "Ŧ", "Ž", "Æ", "Ø", "Å", "Ä", "Ö" ],
- 'sq' => [ "Ç", "Dh", "Ë", "Gj", "Ll", "Nj", "Rr", "Sh", "Th", "Xh", "Zh" ],
- 'sr' => [],
- 'tk' => [ "Ç", "Ä", "Ž", "Ň", "Ö", "Ş", "Ü", "Ý" ],
- 'tl' => [ "Ñ", "Ng" ],
- 'tr' => [ "Ç", "Ğ", "İ", "Ö", "Ş", "Ü" ],
- 'tt' => [ "Ә", "Ө", "Ү", "Җ", "Ң", "Һ" ],
- 'uz' => [ "Ch", "G'", "Ng", "O'", "Sh" ],
- ];
-
- const RECORD_LENGTH = 14;
-
- function __construct( $locale ) {
- if ( !extension_loaded( 'intl' ) ) {
- throw new MWException( 'An ICU collation was requested, ' .
- 'but the intl extension is not available.' );
- }
-
- $this->locale = $locale;
- // Drop everything after the '@' in locale's name
- $localeParts = explode( '@', $locale );
- $this->digitTransformLanguage = Language::factory( $locale === 'root' ? 'en' : $localeParts[0] );
-
- $this->mainCollator = Collator::create( $locale );
- if ( !$this->mainCollator ) {
- throw new MWException( "Invalid ICU locale specified for collation: $locale" );
- }
-
- $this->primaryCollator = Collator::create( $locale );
- $this->primaryCollator->setStrength( Collator::PRIMARY );
- }
-
- function getSortKey( $string ) {
- // intl extension produces non null-terminated
- // strings. Appending '' fixes it so that it doesn't generate
- // a warning on each access in debug php.
- MediaWiki\suppressWarnings();
- $key = $this->mainCollator->getSortKey( $string ) . '';
- MediaWiki\restoreWarnings();
- return $key;
- }
-
- function getPrimarySortKey( $string ) {
- MediaWiki\suppressWarnings();
- $key = $this->primaryCollator->getSortKey( $string ) . '';
- MediaWiki\restoreWarnings();
- return $key;
- }
-
- function getFirstLetter( $string ) {
- $string = strval( $string );
- if ( $string === '' ) {
- return '';
- }
-
- // Check for CJK
- $firstChar = mb_substr( $string, 0, 1, 'UTF-8' );
- if ( ord( $firstChar ) > 0x7f && self::isCjk( UtfNormal\Utils::utf8ToCodepoint( $firstChar ) ) ) {
- return $firstChar;
- }
-
- $sortKey = $this->getPrimarySortKey( $string );
-
- // Do a binary search to find the correct letter to sort under
- $min = ArrayUtils::findLowerBound(
- [ $this, 'getSortKeyByLetterIndex' ],
- $this->getFirstLetterCount(),
- 'strcmp',
- $sortKey );
-
- if ( $min === false ) {
- // Before the first letter
- return '';
- }
- return $this->getLetterByIndex( $min );
- }
-
- function getFirstLetterData() {
- if ( $this->firstLetterData !== null ) {
- return $this->firstLetterData;
- }
-
- $cache = wfGetCache( CACHE_ANYTHING );
- $cacheKey = wfMemcKey(
- 'first-letters',
- $this->locale,
- $this->digitTransformLanguage->getCode(),
- self::getICUVersion()
- );
- $cacheEntry = $cache->get( $cacheKey );
-
- if ( $cacheEntry && isset( $cacheEntry['version'] )
- && $cacheEntry['version'] == self::FIRST_LETTER_VERSION
- ) {
- $this->firstLetterData = $cacheEntry;
- return $this->firstLetterData;
- }
-
- // Generate data from serialized data file
-
- if ( isset( self::$tailoringFirstLetters[$this->locale] ) ) {
- $letters = wfGetPrecompiledData( "first-letters-root.ser" );
- // Append additional characters
- $letters = array_merge( $letters, self::$tailoringFirstLetters[$this->locale] );
- // Remove unnecessary ones, if any
- if ( isset( self::$tailoringFirstLetters['-' . $this->locale] ) ) {
- $letters = array_diff( $letters, self::$tailoringFirstLetters['-' . $this->locale] );
- }
- // Apply digit transforms
- $digits = [ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ];
- $letters = array_diff( $letters, $digits );
- foreach ( $digits as $digit ) {
- $letters[] = $this->digitTransformLanguage->formatNum( $digit, true );
- }
- } else {
- $letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
- if ( $letters === false ) {
- throw new MWException( "MediaWiki does not support ICU locale " .
- "\"{$this->locale}\"" );
- }
- }
-
- /* Sort the letters.
- *
- * It's impossible to have the precompiled data file properly sorted,
- * because the sort order changes depending on ICU version. If the
- * array is not properly sorted, the binary search will return random
- * results.
- *
- * We also take this opportunity to remove primary collisions.
- */
- $letterMap = [];
- foreach ( $letters as $letter ) {
- $key = $this->getPrimarySortKey( $letter );
- if ( isset( $letterMap[$key] ) ) {
- // Primary collision
- // Keep whichever one sorts first in the main collator
- if ( $this->mainCollator->compare( $letter, $letterMap[$key] ) < 0 ) {
- $letterMap[$key] = $letter;
- }
- } else {
- $letterMap[$key] = $letter;
- }
- }
- ksort( $letterMap, SORT_STRING );
-
- /* Remove duplicate prefixes. Basically if something has a sortkey
- * which is a prefix of some other sortkey, then it is an
- * expansion and probably should not be considered a section
- * header.
- *
- * For example 'þ' is sometimes sorted as if it is the letters
- * 'th'. Other times it is its own primary element. Another
- * example is '₨'. Sometimes its a currency symbol. Sometimes it
- * is an 'R' followed by an 's'.
- *
- * Additionally an expanded element should always sort directly
- * after its first element due to they way sortkeys work.
- *
- * UCA sortkey elements are of variable length but no collation
- * element should be a prefix of some other element, so I think
- * this is safe. See:
- * - https://ssl.icu-project.org/repos/icu/icuhtml/trunk/design/collation/ICU_collation_design.htm
- * - http://site.icu-project.org/design/collation/uca-weight-allocation
- *
- * Additionally, there is something called primary compression to
- * worry about. Basically, if you have two primary elements that
- * are more than one byte and both start with the same byte then
- * the first byte is dropped on the second primary. Additionally
- * either \x03 or \xFF may be added to mean that the next primary
- * does not start with the first byte of the first primary.
- *
- * This shouldn't matter much, as the first primary is not
- * changed, and that is what we are comparing against.
- *
- * tl;dr: This makes some assumptions about how icu implements
- * collations. It seems incredibly unlikely these assumptions
- * will change, but nonetheless they are assumptions.
- */
-
- $prev = false;
- $duplicatePrefixes = [];
- foreach ( $letterMap as $key => $value ) {
- // Remove terminator byte. Otherwise the prefix
- // comparison will get hung up on that.
- $trimmedKey = rtrim( $key, "\0" );
- if ( $prev === false || $prev === '' ) {
- $prev = $trimmedKey;
- // We don't yet have a collation element
- // to compare against, so continue.
- continue;
- }
-
- // Due to the fact the array is sorted, we only have
- // to compare with the element directly previous
- // to the current element (skipping expansions).
- // An element "X" will always sort directly
- // before "XZ" (Unless we have "XY", but we
- // do not update $prev in that case).
- if ( substr( $trimmedKey, 0, strlen( $prev ) ) === $prev ) {
- $duplicatePrefixes[] = $key;
- // If this is an expansion, we don't want to
- // compare the next element to this element,
- // but to what is currently $prev
- continue;
- }
- $prev = $trimmedKey;
- }
- foreach ( $duplicatePrefixes as $badKey ) {
- wfDebug( "Removing '{$letterMap[$badKey]}' from first letters.\n" );
- unset( $letterMap[$badKey] );
- // This code assumes that unsetting does not change sort order.
- }
- $data = [
- 'chars' => array_values( $letterMap ),
- 'keys' => array_keys( $letterMap ),
- 'version' => self::FIRST_LETTER_VERSION,
- ];
-
- // Reduce memory usage before caching
- unset( $letterMap );
-
- // Save to cache
- $this->firstLetterData = $data;
- $cache->set( $cacheKey, $data, $cache::TTL_WEEK );
- return $data;
- }
-
- function getLetterByIndex( $index ) {
- if ( $this->firstLetterData === null ) {
- $this->getFirstLetterData();
- }
- return $this->firstLetterData['chars'][$index];
- }
-
- function getSortKeyByLetterIndex( $index ) {
- if ( $this->firstLetterData === null ) {
- $this->getFirstLetterData();
- }
- return $this->firstLetterData['keys'][$index];
- }
-
- function getFirstLetterCount() {
- if ( $this->firstLetterData === null ) {
- $this->getFirstLetterData();
- }
- return count( $this->firstLetterData['chars'] );
- }
-
- static function isCjk( $codepoint ) {
- foreach ( self::$cjkBlocks as $block ) {
- if ( $codepoint >= $block[0] && $codepoint <= $block[1] ) {
- return true;
- }
- }
- return false;
- }
-
- /**
- * Return the version of ICU library used by PHP's intl extension,
- * or false when the extension is not installed of the version
- * can't be determined.
- *
- * The constant INTL_ICU_VERSION this function refers to isn't really
- * documented. It is available since PHP 5.3.7 (see PHP bug 54561).
- * This function will return false on older PHPs.
- *
- * @since 1.21
- * @return string|bool
- */
- static function getICUVersion() {
- return defined( 'INTL_ICU_VERSION' ) ? INTL_ICU_VERSION : false;
- }
-
- /**
- * Return the version of Unicode appropriate for the version of ICU library
- * currently in use, or false when it can't be determined.
- *
- * @since 1.21
- * @return string|bool
- */
- static function getUnicodeVersionForICU() {
- $icuVersion = IcuCollation::getICUVersion();
- if ( !$icuVersion ) {
- return false;
- }
-
- $versionPrefix = substr( $icuVersion, 0, 3 );
- // Source: http://site.icu-project.org/download
- $map = [
- '50.' => '6.2',
- '49.' => '6.1',
- '4.8' => '6.0',
- '4.6' => '6.0',
- '4.4' => '5.2',
- '4.2' => '5.1',
- '4.0' => '5.1',
- '3.8' => '5.0',
- '3.6' => '5.0',
- '3.4' => '4.1',
- ];
-
- if ( isset( $map[$versionPrefix] ) ) {
- return $map[$versionPrefix];
- } else {
- return false;
- }
- }
-}
-
-/**
- * Workaround for the lack of support of Sorani Kurdish / Central Kurdish language ('ckb') in ICU.
- *
- * Uses the same collation rules as Persian / Farsi ('fa'), but different characters for digits.
- */
-class CollationCkb extends IcuCollation {
- function __construct() {
- // This will set $locale and collators, which affect the actual sorting order
- parent::__construct( 'fa' );
- // Override the 'fa' language set by parent constructor, which affects #getFirstLetterData()
- $this->digitTransformLanguage = Language::factory( 'ckb' );
- }
-}
-
-/**
- * Workaround for incorrect collation of Estonian language ('et') in ICU (bug 54168).
- *
- * 'W' and 'V' should not be considered the same letter for the purposes of collation in modern
- * Estonian. We work around this by replacing 'W' and 'w' with 'ᴡ' U+1D21 'LATIN LETTER SMALL
- * CAPITAL W' for sortkey generation, which is collated like 'W' and is not tailored to have the
- * same primary weight as 'V' in Estonian.
- */
-class CollationEt extends IcuCollation {
- function __construct() {
- parent::__construct( 'et' );
- }
-
- private static function mangle( $string ) {
- return str_replace(
- [ 'w', 'W' ],
- 'ᴡ', // U+1D21 'LATIN LETTER SMALL CAPITAL W'
- $string
- );
- }
-
- private static function unmangle( $string ) {
- // Casing data is lost…
- return str_replace(
- 'ᴡ', // U+1D21 'LATIN LETTER SMALL CAPITAL W'
- 'W',
- $string
- );
- }
-
- function getSortKey( $string ) {
- return parent::getSortKey( self::mangle( $string ) );
- }
-
- function getFirstLetter( $string ) {
- return self::unmangle( parent::getFirstLetter( self::mangle( $string ) ) );
- }
-}
/** @var int[] */
private static $pageCount = [];
- /** @var int[] */
- private static $groupMemberCounts = [];
-
static function recache() {
self::load( true );
}
}
if ( get_class( $GLOBALS[$this->global] ) != $this->class ) {
- $fname = __METHOD__ . '-' . $this->global;
$caller = wfGetCaller( $level );
if ( ++$recursionLevel > 2 ) {
throw new MWException( "Unstub loop detected on call of "
*
* @file
*/
+use Cdb\Exception;
+use Cdb\Reader;
+use Cdb\Writer;
/**
* LCStore implementation which stores data as a collection of CDB files in the
*/
class LCStoreCDB implements LCStore {
- /** @var CdbReader[] */
+ /** @var Reader[] */
private $readers;
- /** @var CdbWriter */
+ /** @var Writer */
private $writer;
/** @var string Current language code */
$this->readers[$code] = false;
if ( file_exists( $fileName ) ) {
try {
- $this->readers[$code] = CdbReader::open( $fileName );
- } catch ( CdbException $e ) {
+ $this->readers[$code] = Reader::open( $fileName );
+ } catch ( Exception $e ) {
wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
}
}
$value = false;
try {
$value = $this->readers[$code]->get( $key );
- } catch ( CdbException $e ) {
- wfDebug( __METHOD__ . ": CdbException caught, error message was "
+ } catch ( Exception $e ) {
+ wfDebug( __METHOD__ . ": \Cdb\Exception caught, error message was "
. $e->getMessage() . "\n" );
}
if ( $value === false ) {
}
try {
- $this->writer = CdbWriter::open( $this->getFileName( $code ) );
- } catch ( CdbException $e ) {
+ $this->writer = Writer::open( $this->getFileName( $code ) );
+ } catch ( Exception $e ) {
throw new MWException( $e->getMessage() );
}
$this->currentLang = $code;
// Close the writer
try {
$this->writer->close();
- } catch ( CdbException $e ) {
+ } catch ( Exception $e ) {
throw new MWException( $e->getMessage() );
}
$this->writer = null;
}
try {
$this->writer->set( $key, serialize( $value ) );
- } catch ( CdbException $e ) {
+ } catch ( Exception $e ) {
throw new MWException( $e->getMessage() );
}
}
--- /dev/null
+<?php
+/**
+ * Database row sorting.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @file
+ */
+
+/**
+ * @since 1.16.3
+ * @author Tim Starling
+ */
+abstract class Collation {
+ private static $instance;
+
+ /**
+ * @since 1.16.3
+ * @return Collation
+ */
+ public static function singleton() {
+ if ( !self::$instance ) {
+ global $wgCategoryCollation;
+ self::$instance = self::factory( $wgCategoryCollation );
+ }
+ return self::$instance;
+ }
+
+ /**
+ * @since 1.16.3
+ * @throws MWException
+ * @param string $collationName
+ * @return Collation
+ */
+ public static function factory( $collationName ) {
+ switch ( $collationName ) {
+ case 'uppercase':
+ return new UppercaseCollation;
+ case 'identity':
+ return new IdentityCollation;
+ case 'uca-default':
+ return new IcuCollation( 'root' );
+ case 'xx-uca-ckb':
+ return new CollationCkb;
+ case 'xx-uca-et':
+ return new CollationEt;
+ default:
+ $match = [];
+ if ( preg_match( '/^uca-([a-z@=-]+)$/', $collationName, $match ) ) {
+ return new IcuCollation( $match[1] );
+ }
+
+ # Provide a mechanism for extensions to hook in.
+ $collationObject = null;
+ Hooks::run( 'Collation::factory', [ $collationName, &$collationObject ] );
+
+ if ( $collationObject instanceof Collation ) {
+ return $collationObject;
+ }
+
+ // If all else fails...
+ throw new MWException( __METHOD__ . ": unknown collation type \"$collationName\"" );
+ }
+ }
+
+ /**
+ * Given a string, convert it to a (hopefully short) key that can be used
+ * for efficient sorting. A binary sort according to the sortkeys
+ * corresponds to a logical sort of the corresponding strings. Current
+ * code expects that a line feed character should sort before all others, but
+ * has no other particular expectations (and that one can be changed if
+ * necessary).
+ *
+ * @since 1.16.3
+ *
+ * @param string $string UTF-8 string
+ * @return string Binary sortkey
+ */
+ abstract function getSortKey( $string );
+
+ /**
+ * Given a string, return the logical "first letter" to be used for
+ * grouping on category pages and so on. This has to be coordinated
+ * carefully with convertToSortkey(), or else the sorted list might jump
+ * back and forth between the same "initial letters" or other pathological
+ * behavior. For instance, if you just return the first character, but "a"
+ * sorts the same as "A" based on getSortKey(), then you might get a
+ * list like
+ *
+ * == A ==
+ * * [[Aardvark]]
+ *
+ * == a ==
+ * * [[antelope]]
+ *
+ * == A ==
+ * * [[Ape]]
+ *
+ * etc., assuming for the sake of argument that $wgCapitalLinks is false.
+ *
+ * @since 1.16.3
+ *
+ * @param string $string UTF-8 string
+ * @return string UTF-8 string corresponding to the first letter of input
+ */
+ abstract function getFirstLetter( $string );
+
+}
--- /dev/null
+<?php
+/**
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @file
+ */
+
+/**
+ * Workaround for the lack of support of Sorani Kurdish / Central Kurdish language ('ckb') in ICU.
+ *
+ * Uses the same collation rules as Persian / Farsi ('fa'), but different characters for digits.
+ *
+ * @since 1.23
+ */
+class CollationCkb extends IcuCollation {
+ public function __construct() {
+ // This will set $locale and collators, which affect the actual sorting order
+ parent::__construct( 'fa' );
+ // Override the 'fa' language set by parent constructor, which affects #getFirstLetterData()
+ $this->digitTransformLanguage = Language::factory( 'ckb' );
+ }
+}
--- /dev/null
+<?php
+/**
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @file
+ */
+
+/**
+ * Workaround for incorrect collation of Estonian language ('et') in ICU (bug 54168).
+ *
+ * 'W' and 'V' should not be considered the same letter for the purposes of collation in modern
+ * Estonian. We work around this by replacing 'W' and 'w' with 'ᴡ' U+1D21 'LATIN LETTER SMALL
+ * CAPITAL W' for sortkey generation, which is collated like 'W' and is not tailored to have the
+ * same primary weight as 'V' in Estonian.
+ *
+ * @since 1.24
+ */
+class CollationEt extends IcuCollation {
+ public function __construct() {
+ parent::__construct( 'et' );
+ }
+
+ private static function mangle( $string ) {
+ return str_replace(
+ [ 'w', 'W' ],
+ 'ᴡ', // U+1D21 'LATIN LETTER SMALL CAPITAL W'
+ $string
+ );
+ }
+
+ private static function unmangle( $string ) {
+ // Casing data is lost…
+ return str_replace(
+ 'ᴡ', // U+1D21 'LATIN LETTER SMALL CAPITAL W'
+ 'W',
+ $string
+ );
+ }
+
+ public function getSortKey( $string ) {
+ return parent::getSortKey( self::mangle( $string ) );
+ }
+
+ public function getFirstLetter( $string ) {
+ return self::unmangle( parent::getFirstLetter( self::mangle( $string ) ) );
+ }
+}
--- /dev/null
+<?php
+/**
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @file
+ */
+
+/**
+ * @since 1.16.3
+ */
+class IcuCollation extends Collation {
+ const FIRST_LETTER_VERSION = 2;
+
+ /** @var Collator */
+ private $primaryCollator;
+
+ /** @var Collator */
+ private $mainCollator;
+
+ /** @var string */
+ private $locale;
+
+ /** @var Language */
+ protected $digitTransformLanguage;
+
+ /** @var array */
+ private $firstLetterData;
+
+ /**
+ * Unified CJK blocks.
+ *
+ * The same definition of a CJK block must be used for both Collation and
+ * generateCollationData.php. These blocks are omitted from the first
+ * letter data, as an optimisation measure and because the default UCA table
+ * is pretty useless for sorting Chinese text anyway. Japanese and Korean
+ * blocks are not included here, because they are smaller and more useful.
+ */
+ private static $cjkBlocks = [
+ [ 0x2E80, 0x2EFF ], // CJK Radicals Supplement
+ [ 0x2F00, 0x2FDF ], // Kangxi Radicals
+ [ 0x2FF0, 0x2FFF ], // Ideographic Description Characters
+ [ 0x3000, 0x303F ], // CJK Symbols and Punctuation
+ [ 0x31C0, 0x31EF ], // CJK Strokes
+ [ 0x3200, 0x32FF ], // Enclosed CJK Letters and Months
+ [ 0x3300, 0x33FF ], // CJK Compatibility
+ [ 0x3400, 0x4DBF ], // CJK Unified Ideographs Extension A
+ [ 0x4E00, 0x9FFF ], // CJK Unified Ideographs
+ [ 0xF900, 0xFAFF ], // CJK Compatibility Ideographs
+ [ 0xFE30, 0xFE4F ], // CJK Compatibility Forms
+ [ 0x20000, 0x2A6DF ], // CJK Unified Ideographs Extension B
+ [ 0x2A700, 0x2B73F ], // CJK Unified Ideographs Extension C
+ [ 0x2B740, 0x2B81F ], // CJK Unified Ideographs Extension D
+ [ 0x2F800, 0x2FA1F ], // CJK Compatibility Ideographs Supplement
+ ];
+
+ /**
+ * Additional characters (or character groups) to be considered separate
+ * letters for given languages, or to be removed from the list of such
+ * letters (denoted by keys starting with '-').
+ *
+ * These are additions to (or subtractions from) the data stored in the
+ * first-letters-root.ser file (which among others includes full basic latin,
+ * cyrillic and greek alphabets).
+ *
+ * "Separate letter" is a letter that would have a separate heading/section
+ * for it in a dictionary or a phone book in this language. This data isn't
+ * used for sorting (the ICU library handles that), only for deciding which
+ * characters (or character groups) to use as headings.
+ *
+ * Initially generated based on the primary level of Unicode collation
+ * tailorings available at http://developer.mimer.com/charts/tailorings.htm ,
+ * later modified.
+ *
+ * Empty arrays are intended; this signifies that the data for the language is
+ * available and that there are, in fact, no additional letters to consider.
+ */
+ private static $tailoringFirstLetters = [
+ // Verified by native speakers
+ 'be' => [ "Ё" ],
+ 'be-tarask' => [ "Ё" ],
+ 'cy' => [ "Ch", "Dd", "Ff", "Ng", "Ll", "Ph", "Rh", "Th" ],
+ 'en' => [],
+ 'fa' => [ "آ", "ء", "ه" ],
+ 'fi' => [ "Å", "Ä", "Ö" ],
+ 'fr' => [],
+ 'hu' => [ "Cs", "Dz", "Dzs", "Gy", "Ly", "Ny", "Ö", "Sz", "Ty", "Ü", "Zs" ],
+ 'is' => [ "Á", "Ð", "É", "Í", "Ó", "Ú", "Ý", "Þ", "Æ", "Ö", "Å" ],
+ 'it' => [],
+ 'lv' => [ "Č", "Ģ", "Ķ", "Ļ", "Ņ", "Š", "Ž" ],
+ 'pl' => [ "Ą", "Ć", "Ę", "Ł", "Ń", "Ó", "Ś", "Ź", "Ż" ],
+ 'pt' => [],
+ 'ru' => [],
+ 'sv' => [ "Å", "Ä", "Ö" ],
+ 'sv@collation=standard' => [ "Å", "Ä", "Ö" ],
+ 'uk' => [ "Ґ", "Ь" ],
+ 'vi' => [ "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ],
+ // Not verified, but likely correct
+ 'af' => [],
+ 'ast' => [ "Ch", "Ll", "Ñ" ],
+ 'az' => [ "Ç", "Ə", "Ğ", "İ", "Ö", "Ş", "Ü" ],
+ 'bg' => [],
+ 'br' => [ "Ch", "C'h" ],
+ 'bs' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ],
+ 'ca' => [],
+ 'co' => [],
+ 'cs' => [ "Č", "Ch", "Ř", "Š", "Ž" ],
+ 'da' => [ "Æ", "Ø", "Å" ],
+ 'de' => [],
+ 'dsb' => [ "Č", "Ć", "Dź", "Ě", "Ch", "Ł", "Ń", "Ŕ", "Š", "Ś", "Ž", "Ź" ],
+ 'el' => [],
+ 'eo' => [ "Ĉ", "Ĝ", "Ĥ", "Ĵ", "Ŝ", "Ŭ" ],
+ 'es' => [ "Ñ" ],
+ 'et' => [ "Š", "Ž", "Õ", "Ä", "Ö", "Ü", "W" ], // added W for CollationEt (xx-uca-et)
+ 'eu' => [ "Ñ" ],
+ 'fo' => [ "Á", "Ð", "Í", "Ó", "Ú", "Ý", "Æ", "Ø", "Å" ],
+ 'fur' => [ "À", "Á", "Â", "È", "Ì", "Ò", "Ù" ],
+ 'fy' => [],
+ 'ga' => [],
+ 'gd' => [],
+ 'gl' => [ "Ch", "Ll", "Ñ" ],
+ 'hr' => [ "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ],
+ 'hsb' => [ "Č", "Dź", "Ě", "Ch", "Ł", "Ń", "Ř", "Š", "Ć", "Ž" ],
+ 'kk' => [ "Ү", "І" ],
+ 'kl' => [ "Æ", "Ø", "Å" ],
+ 'ku' => [ "Ç", "Ê", "Î", "Ş", "Û" ],
+ 'ky' => [ "Ё" ],
+ 'la' => [],
+ 'lb' => [],
+ 'lt' => [ "Č", "Š", "Ž" ],
+ 'mk' => [],
+ 'mo' => [ "Ă", "Â", "Î", "Ş", "Ţ" ],
+ 'mt' => [ "Ċ", "Ġ", "Għ", "Ħ", "Ż" ],
+ 'nl' => [],
+ 'no' => [ "Æ", "Ø", "Å" ],
+ 'oc' => [],
+ 'rm' => [],
+ 'ro' => [ "Ă", "Â", "Î", "Ş", "Ţ" ],
+ 'rup' => [ "Ă", "Â", "Î", "Ľ", "Ń", "Ş", "Ţ" ],
+ 'sco' => [],
+ 'sk' => [ "Ä", "Č", "Ch", "Ô", "Š", "Ž" ],
+ 'sl' => [ "Č", "Š", "Ž" ],
+ 'smn' => [ "Á", "Č", "Đ", "Ŋ", "Š", "Ŧ", "Ž", "Æ", "Ø", "Å", "Ä", "Ö" ],
+ 'sq' => [ "Ç", "Dh", "Ë", "Gj", "Ll", "Nj", "Rr", "Sh", "Th", "Xh", "Zh" ],
+ 'sr' => [],
+ 'tk' => [ "Ç", "Ä", "Ž", "Ň", "Ö", "Ş", "Ü", "Ý" ],
+ 'tl' => [ "Ñ", "Ng" ],
+ 'tr' => [ "Ç", "Ğ", "İ", "Ö", "Ş", "Ü" ],
+ 'tt' => [ "Ә", "Ө", "Ү", "Җ", "Ң", "Һ" ],
+ 'uz' => [ "Ch", "G'", "Ng", "O'", "Sh" ],
+ ];
+
+ /**
+ * @since 1.16.3
+ */
+ const RECORD_LENGTH = 14;
+
+ public function __construct( $locale ) {
+ if ( !extension_loaded( 'intl' ) ) {
+ throw new MWException( 'An ICU collation was requested, ' .
+ 'but the intl extension is not available.' );
+ }
+
+ $this->locale = $locale;
+ // Drop everything after the '@' in locale's name
+ $localeParts = explode( '@', $locale );
+ $this->digitTransformLanguage = Language::factory( $locale === 'root' ? 'en' : $localeParts[0] );
+
+ $this->mainCollator = Collator::create( $locale );
+ if ( !$this->mainCollator ) {
+ throw new MWException( "Invalid ICU locale specified for collation: $locale" );
+ }
+
+ $this->primaryCollator = Collator::create( $locale );
+ $this->primaryCollator->setStrength( Collator::PRIMARY );
+ }
+
+ public function getSortKey( $string ) {
+ // intl extension produces non null-terminated
+ // strings. Appending '' fixes it so that it doesn't generate
+ // a warning on each access in debug php.
+ MediaWiki\suppressWarnings();
+ $key = $this->mainCollator->getSortKey( $string ) . '';
+ MediaWiki\restoreWarnings();
+ return $key;
+ }
+
+ public function getPrimarySortKey( $string ) {
+ MediaWiki\suppressWarnings();
+ $key = $this->primaryCollator->getSortKey( $string ) . '';
+ MediaWiki\restoreWarnings();
+ return $key;
+ }
+
+ public function getFirstLetter( $string ) {
+ $string = strval( $string );
+ if ( $string === '' ) {
+ return '';
+ }
+
+ // Check for CJK
+ $firstChar = mb_substr( $string, 0, 1, 'UTF-8' );
+ if ( ord( $firstChar ) > 0x7f && self::isCjk( UtfNormal\Utils::utf8ToCodepoint( $firstChar ) ) ) {
+ return $firstChar;
+ }
+
+ $sortKey = $this->getPrimarySortKey( $string );
+
+ // Do a binary search to find the correct letter to sort under
+ $min = ArrayUtils::findLowerBound(
+ [ $this, 'getSortKeyByLetterIndex' ],
+ $this->getFirstLetterCount(),
+ 'strcmp',
+ $sortKey );
+
+ if ( $min === false ) {
+ // Before the first letter
+ return '';
+ }
+ return $this->getLetterByIndex( $min );
+ }
+
+ /**
+ * @since 1.16.3
+ */
+ public function getFirstLetterData() {
+ if ( $this->firstLetterData !== null ) {
+ return $this->firstLetterData;
+ }
+
+ $cache = wfGetCache( CACHE_ANYTHING );
+ $cacheKey = wfMemcKey(
+ 'first-letters',
+ $this->locale,
+ $this->digitTransformLanguage->getCode(),
+ self::getICUVersion()
+ );
+ $cacheEntry = $cache->get( $cacheKey );
+
+ if ( $cacheEntry && isset( $cacheEntry['version'] )
+ && $cacheEntry['version'] == self::FIRST_LETTER_VERSION
+ ) {
+ $this->firstLetterData = $cacheEntry;
+ return $this->firstLetterData;
+ }
+
+ // Generate data from serialized data file
+
+ if ( isset( self::$tailoringFirstLetters[$this->locale] ) ) {
+ $letters = wfGetPrecompiledData( "first-letters-root.ser" );
+ // Append additional characters
+ $letters = array_merge( $letters, self::$tailoringFirstLetters[$this->locale] );
+ // Remove unnecessary ones, if any
+ if ( isset( self::$tailoringFirstLetters['-' . $this->locale] ) ) {
+ $letters = array_diff( $letters, self::$tailoringFirstLetters['-' . $this->locale] );
+ }
+ // Apply digit transforms
+ $digits = [ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ];
+ $letters = array_diff( $letters, $digits );
+ foreach ( $digits as $digit ) {
+ $letters[] = $this->digitTransformLanguage->formatNum( $digit, true );
+ }
+ } else {
+ $letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
+ if ( $letters === false ) {
+ throw new MWException( "MediaWiki does not support ICU locale " .
+ "\"{$this->locale}\"" );
+ }
+ }
+
+ /* Sort the letters.
+ *
+ * It's impossible to have the precompiled data file properly sorted,
+ * because the sort order changes depending on ICU version. If the
+ * array is not properly sorted, the binary search will return random
+ * results.
+ *
+ * We also take this opportunity to remove primary collisions.
+ */
+ $letterMap = [];
+ foreach ( $letters as $letter ) {
+ $key = $this->getPrimarySortKey( $letter );
+ if ( isset( $letterMap[$key] ) ) {
+ // Primary collision
+ // Keep whichever one sorts first in the main collator
+ if ( $this->mainCollator->compare( $letter, $letterMap[$key] ) < 0 ) {
+ $letterMap[$key] = $letter;
+ }
+ } else {
+ $letterMap[$key] = $letter;
+ }
+ }
+ ksort( $letterMap, SORT_STRING );
+
+ /* Remove duplicate prefixes. Basically if something has a sortkey
+ * which is a prefix of some other sortkey, then it is an
+ * expansion and probably should not be considered a section
+ * header.
+ *
+ * For example 'þ' is sometimes sorted as if it is the letters
+ * 'th'. Other times it is its own primary element. Another
+ * example is '₨'. Sometimes its a currency symbol. Sometimes it
+ * is an 'R' followed by an 's'.
+ *
+ * Additionally an expanded element should always sort directly
+ * after its first element due to they way sortkeys work.
+ *
+ * UCA sortkey elements are of variable length but no collation
+ * element should be a prefix of some other element, so I think
+ * this is safe. See:
+ * - https://ssl.icu-project.org/repos/icu/icuhtml/trunk/design/collation/ICU_collation_design.htm
+ * - http://site.icu-project.org/design/collation/uca-weight-allocation
+ *
+ * Additionally, there is something called primary compression to
+ * worry about. Basically, if you have two primary elements that
+ * are more than one byte and both start with the same byte then
+ * the first byte is dropped on the second primary. Additionally
+ * either \x03 or \xFF may be added to mean that the next primary
+ * does not start with the first byte of the first primary.
+ *
+ * This shouldn't matter much, as the first primary is not
+ * changed, and that is what we are comparing against.
+ *
+ * tl;dr: This makes some assumptions about how icu implements
+ * collations. It seems incredibly unlikely these assumptions
+ * will change, but nonetheless they are assumptions.
+ */
+
+ $prev = false;
+ $duplicatePrefixes = [];
+ foreach ( $letterMap as $key => $value ) {
+ // Remove terminator byte. Otherwise the prefix
+ // comparison will get hung up on that.
+ $trimmedKey = rtrim( $key, "\0" );
+ if ( $prev === false || $prev === '' ) {
+ $prev = $trimmedKey;
+ // We don't yet have a collation element
+ // to compare against, so continue.
+ continue;
+ }
+
+ // Due to the fact the array is sorted, we only have
+ // to compare with the element directly previous
+ // to the current element (skipping expansions).
+ // An element "X" will always sort directly
+ // before "XZ" (Unless we have "XY", but we
+ // do not update $prev in that case).
+ if ( substr( $trimmedKey, 0, strlen( $prev ) ) === $prev ) {
+ $duplicatePrefixes[] = $key;
+ // If this is an expansion, we don't want to
+ // compare the next element to this element,
+ // but to what is currently $prev
+ continue;
+ }
+ $prev = $trimmedKey;
+ }
+ foreach ( $duplicatePrefixes as $badKey ) {
+ wfDebug( "Removing '{$letterMap[$badKey]}' from first letters.\n" );
+ unset( $letterMap[$badKey] );
+ // This code assumes that unsetting does not change sort order.
+ }
+ $data = [
+ 'chars' => array_values( $letterMap ),
+ 'keys' => array_keys( $letterMap ),
+ 'version' => self::FIRST_LETTER_VERSION,
+ ];
+
+ // Reduce memory usage before caching
+ unset( $letterMap );
+
+ // Save to cache
+ $this->firstLetterData = $data;
+ $cache->set( $cacheKey, $data, $cache::TTL_WEEK );
+ return $data;
+ }
+
+ /**
+ * @since 1.16.3
+ */
+ public function getLetterByIndex( $index ) {
+ if ( $this->firstLetterData === null ) {
+ $this->getFirstLetterData();
+ }
+ return $this->firstLetterData['chars'][$index];
+ }
+
+ /**
+ * @since 1.16.3
+ */
+ public function getSortKeyByLetterIndex( $index ) {
+ if ( $this->firstLetterData === null ) {
+ $this->getFirstLetterData();
+ }
+ return $this->firstLetterData['keys'][$index];
+ }
+
+ /**
+ * @since 1.16.3
+ */
+ public function getFirstLetterCount() {
+ if ( $this->firstLetterData === null ) {
+ $this->getFirstLetterData();
+ }
+ return count( $this->firstLetterData['chars'] );
+ }
+
+ /**
+ * @since 1.16.3
+ */
+ public static function isCjk( $codepoint ) {
+ foreach ( self::$cjkBlocks as $block ) {
+ if ( $codepoint >= $block[0] && $codepoint <= $block[1] ) {
+ return true;
+ }
+ }
+ return false;
+ }
+
+ /**
+ * Return the version of ICU library used by PHP's intl extension,
+ * or false when the extension is not installed of the version
+ * can't be determined.
+ *
+ * The constant INTL_ICU_VERSION this function refers to isn't really
+ * documented. It is available since PHP 5.3.7 (see PHP bug 54561).
+ * This function will return false on older PHPs.
+ *
+ * @since 1.21
+ * @return string|bool
+ */
+ static function getICUVersion() {
+ return defined( 'INTL_ICU_VERSION' ) ? INTL_ICU_VERSION : false;
+ }
+
+ /**
+ * Return the version of Unicode appropriate for the version of ICU library
+ * currently in use, or false when it can't be determined.
+ *
+ * @since 1.21
+ * @return string|bool
+ */
+ static function getUnicodeVersionForICU() {
+ $icuVersion = IcuCollation::getICUVersion();
+ if ( !$icuVersion ) {
+ return false;
+ }
+
+ $versionPrefix = substr( $icuVersion, 0, 3 );
+ // Source: http://site.icu-project.org/download
+ $map = [
+ '50.' => '6.2',
+ '49.' => '6.1',
+ '4.8' => '6.0',
+ '4.6' => '6.0',
+ '4.4' => '5.2',
+ '4.2' => '5.1',
+ '4.0' => '5.1',
+ '3.8' => '5.0',
+ '3.6' => '5.0',
+ '3.4' => '4.1',
+ ];
+
+ if ( isset( $map[$versionPrefix] ) ) {
+ return $map[$versionPrefix];
+ } else {
+ return false;
+ }
+ }
+}
--- /dev/null
+<?php
+/**
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @file
+ */
+
+/**
+ * Collation class that's essentially a no-op.
+ *
+ * Does sorting based on binary value of the string.
+ * Like how things were pre 1.17.
+ *
+ * @since 1.18
+ */
+class IdentityCollation extends Collation {
+
+ public function getSortKey( $string ) {
+ return $string;
+ }
+
+ public function getFirstLetter( $string ) {
+ global $wgContLang;
+ // Copied from UppercaseCollation.
+ // I'm kind of unclear on when this could happen...
+ if ( $string[0] == "\0" ) {
+ $string = substr( $string, 1 );
+ }
+ return $wgContLang->firstChar( $string );
+ }
+}
--- /dev/null
+<?php
+/**
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @since 1.16.3
+ *
+ * @file
+ */
+
+class UppercaseCollation extends Collation {
+
+ private $lang;
+
+ public function __construct() {
+ // Get a language object so that we can use the generic UTF-8 uppercase
+ // function there
+ $this->lang = Language::factory( 'en' );
+ }
+
+ public function getSortKey( $string ) {
+ return $this->lang->uc( $string );
+ }
+
+ public function getFirstLetter( $string ) {
+ if ( $string[0] == "\0" ) {
+ $string = substr( $string, 1 );
+ }
+ return $this->lang->ucfirst( $this->lang->firstChar( $string ) );
+ }
+
+}
abstract protected function doNotify( $channel, array $events );
}
-/**
- * No-op class for publishing messages into a PubSub system
- */
-class EventRelayerNull extends EventRelayer {
- public function doNotify( $channel, array $events ) {
- return true;
- }
-}
--- /dev/null
+<?php
+/**
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ * http://www.gnu.org/copyleft/gpl.html
+ *
+ * @file
+ * @author Aaron Schulz
+ */
+
+/**
+ * No-op class for publishing messages into a PubSub system
+ */
+class EventRelayerNull extends EventRelayer {
+ public function doNotify( $channel, array $events ) {
+ return true;
+ }
+}
protected function transformImageMagick( $image, $params ) {
# use ImageMagick
global $wgSharpenReductionThreshold, $wgSharpenParameter, $wgMaxAnimatedGifArea,
- $wgImageMagickTempDir, $wgImageMagickConvertCommand, $wgMaxInterlacingAreas;
+ $wgImageMagickTempDir, $wgImageMagickConvertCommand;
$quality = [];
$sharpen = [];
return [];
}
- $search = SearchEngine::create();
- $search->setLimitOffset( $limit, $offset );
- $search->setNamespaces( [] );
- $result = $search->defaultPrefixSearch( $search );
+ $searchEngine = SearchEngine::create();
+ $searchEngine->setLimitOffset( $limit, $offset );
+ $searchEngine->setNamespaces( [] );
+ $result = $searchEngine->defaultPrefixSearch( $search );
return array_map( function( Title $t ) {
return $t->getPrefixedText();
}, $result );
}
protected function getFormFields() {
- $user = $this->getUser();
- $request = $this->getRequest();
-
$fields = [];
if ( $this->par !== null ) {
}
public function execute( $par ) {
- $miserMode = $this->getConfig()->get( 'MiserMode' );
-
$this->setHeaders();
$this->getOutput()->addModuleStyles( 'mediawiki.special' );
$server # output channel
);
- $mcc = new MemCachedClientforWiki( [
+ $mcc = new MemcachedClient( [
'persistant' => true,
'timeout' => $wgMemCachedTimeout
] );