X-Git-Url: https://git.heureux-cyclage.org/?a=blobdiff_plain;f=includes%2FLinksUpdate.php;h=ef3374d9f962c954683e67eaea76b97b0309c2b5;hb=c7a8875329494a1f184049c670c05c518af742b2;hp=f909c35d9794b5379f9be1f19ef2a906802c4043;hpb=eb53cc08560721208e195c0f073809e7b3eee485;p=lhc%2Fweb%2Fwiklou.git diff --git a/includes/LinksUpdate.php b/includes/LinksUpdate.php index f909c35d97..ef3374d9f9 100644 --- a/includes/LinksUpdate.php +++ b/includes/LinksUpdate.php @@ -1,37 +1,36 @@ mOptions = array( 'FOR UPDATE' ); } - $this->mDb =& wfGetDB( DB_MASTER ); + $this->mDb = wfGetDB( DB_MASTER ); if ( !is_object( $title ) ) { - wfDebugDieBacktrace( "The calling convention to LinksUpdate::LinksUpdate() has changed. " . + throw new MWException( "The calling convention to LinksUpdate::LinksUpdate() has changed. " . "Please see Article::editUpdates() for an invocation example.\n" ); } $this->mTitle = $title; $this->mId = $title->getArticleID(); + $this->mParserOutput = $parserOutput; + $this->mLinks = $parserOutput->getLinks(); + $this->mImages = $parserOutput->getImages(); + $this->mTemplates = $parserOutput->getTemplates(); + $this->mExternals = $parserOutput->getExternalLinks(); + $this->mCategories = $parserOutput->getCategories(); + $this->mProperties = $parserOutput->getProperties(); + + # Convert the format of the interlanguage links + # I didn't want to change it in the ParserOutput, because that array is passed all + # the way back to the skin, so either a skin API break would be required, or an + # inefficient back-conversion. + $ill = $parserOutput->getLanguageLinks(); + $this->mInterlangs = array(); + foreach ( $ill as $link ) { + list( $key, $title ) = explode( ':', $link, 2 ); + $this->mInterlangs[$key] = $title; + } - // Shortcut aliases - $this->mLinks =& $this->mParserOutput->getLinks(); - $this->mImages =& $this->mParserOutput->getImages(); - $this->mTemplates =& $this->mParserOutput->getTemplates(); - $this->mExternals =& $this->mParserOutput->getExternalLinks(); - $this->mCategories =& $this->mParserOutput->getCategories(); + $this->mRecursive = $recursive; + $this->mTouchTmplLinks = false; + wfRunHooks( 'LinksUpdateConstructed', array( &$this ) ); } /** * Update link tables with outgoing links from an updated article */ - function doUpdate() { + public function doUpdate() { global $wgUseDumbLinkUpdate; + + wfRunHooks( 'LinksUpdate', array( &$this ) ); if ( $wgUseDumbLinkUpdate ) { $this->doDumbUpdate(); } else { $this->doIncrementalUpdate(); } + wfRunHooks( 'LinksUpdateComplete', array( &$this ) ); } - function doIncrementalUpdate() { - $fname = 'LinksUpdate::doIncrementalUpdate'; - wfProfileIn( $fname ); + protected function doIncrementalUpdate() { + wfProfileIn( __METHOD__ ); # Page links $existing = $this->getExistingLinks(); $this->incrTableUpdate( 'pagelinks', 'pl', $this->getLinkDeletions( $existing ), $this->getLinkInsertions( $existing ) ); - # Template links - $existing = $this->getExistingTemplates(); - $this->incrTableUpdate( 'templatelinks', 'tl', $this->getTemplateDeletions( $existing ), - $this->getTemplateInsertions( $existing ) ); - # Image links $existing = $this->getExistingImages(); - $this->incrTableUpdate( 'imagelinks', 'il', $this->getImageDeletions( $existing ), - $this->getImageInsertions( $existing ) ); + + $imageDeletes = $this->getImageDeletions( $existing ); + $this->incrTableUpdate( 'imagelinks', 'il', $imageDeletes, $this->getImageInsertions( $existing ) ); + + # Invalidate all image description pages which had links added or removed + $imageUpdates = $imageDeletes + array_diff_key( $this->mImages, $existing ); + $this->invalidateImageDescriptions( $imageUpdates ); # External links $existing = $this->getExistingExternals(); $this->incrTableUpdate( 'externallinks', 'el', $this->getExternalDeletions( $existing ), - $this->getExternalInsertions( $existing ) ); - + $this->getExternalInsertions( $existing ) ); + + # Language links + $existing = $this->getExistingInterlangs(); + $this->incrTableUpdate( 'langlinks', 'll', $this->getInterlangDeletions( $existing ), + $this->getInterlangInsertions( $existing ) ); + + # Template links + $existing = $this->getExistingTemplates(); + $this->incrTableUpdate( 'templatelinks', 'tl', $this->getTemplateDeletions( $existing ), + $this->getTemplateInsertions( $existing ) ); + # Category links $existing = $this->getExistingCategories(); - $this->incrTableUpdate( 'categorylinks', 'cl', $this->getCategoryDeletions( $existing ), - $this->getCategoryInsertions( $existing ) ); - # I think this works out to a set XOR operation, the idea is to invalidate all - # categories which were added, deleted or changed - # FIXME: surely there's a more appropriate place to put this update? - $categoryUpdates = array_diff_assoc( $existing, $this->mCategories ) + array_diff_assoc( $this->mCategories, $existing ); + $categoryDeletes = $this->getCategoryDeletions( $existing ); + + $this->incrTableUpdate( 'categorylinks', 'cl', $categoryDeletes, $this->getCategoryInsertions( $existing ) ); + + # Invalidate all categories which were added, deleted or changed (set symmetric difference) + $categoryInserts = array_diff_assoc( $this->mCategories, $existing ); + $categoryUpdates = $categoryInserts + $categoryDeletes; $this->invalidateCategories( $categoryUpdates ); + $this->updateCategoryCounts( $categoryInserts, $categoryDeletes ); + + # Page properties + $existing = $this->getExistingProperties(); + + $propertiesDeletes = $this->getPropertyDeletions( $existing ); + + $this->incrTableUpdate( 'page_props', 'pp', $propertiesDeletes, $this->getPropertyInsertions( $existing ) ); + + # Invalidate the necessary pages + $changed = $propertiesDeletes + array_diff_assoc( $this->mProperties, $existing ); + $this->invalidateProperties( $changed ); + + # Refresh links of all pages including this page + # This will be in a separate transaction + if ( $this->mRecursive ) { + $this->queueRecursiveJobs(); + } - wfProfileOut( $fname ); + wfProfileOut( __METHOD__ ); } /** - * Link update which clears the previous entries and inserts new ones - * May be slower or faster depending on level of lock contention and write speed of DB - * Also useful where link table corruption needs to be repaired, e.g. in refreshLinks.php + * Link update which clears the previous entries and inserts new ones + * May be slower or faster depending on level of lock contention and write speed of DB + * Also useful where link table corruption needs to be repaired, e.g. in refreshLinks.php */ - function doDumbUpdate() { - $fname = 'LinksUpdate::doDumbUpdate'; - wfProfileIn( $fname ); + protected function doDumbUpdate() { + wfProfileIn( __METHOD__ ); + # Refresh category pages and image description pages $existing = $this->getExistingCategories(); - $categoryUpdates = array_diff_assoc( $existing, $this->mCategories ) + array_diff_assoc( $this->mCategories, $existing ); + $categoryInserts = array_diff_assoc( $this->mCategories, $existing ); + $categoryDeletes = array_diff_assoc( $existing, $this->mCategories ); + $categoryUpdates = $categoryInserts + $categoryDeletes; + $existing = $this->getExistingImages(); + $imageUpdates = array_diff_key( $existing, $this->mImages ) + array_diff_key( $this->mImages, $existing ); $this->dumbTableUpdate( 'pagelinks', $this->getLinkInsertions(), 'pl_from' ); $this->dumbTableUpdate( 'imagelinks', $this->getImageInsertions(), 'il_from' ); $this->dumbTableUpdate( 'categorylinks', $this->getCategoryInsertions(), 'cl_from' ); $this->dumbTableUpdate( 'templatelinks', $this->getTemplateInsertions(), 'tl_from' ); $this->dumbTableUpdate( 'externallinks', $this->getExternalInsertions(), 'el_from' ); + $this->dumbTableUpdate( 'langlinks', $this->getInterlangInsertions(),'ll_from' ); + $this->dumbTableUpdate( 'page_props', $this->getPropertyInsertions(), 'pp_page' ); - # Update the cache of all the category pages + # Update the cache of all the category pages and image description + # pages which were changed, and fix the category table count $this->invalidateCategories( $categoryUpdates ); + $this->updateCategoryCounts( $categoryInserts, $categoryDeletes ); + $this->invalidateImageDescriptions( $imageUpdates ); + + # Refresh links of all pages including this page + # This will be in a separate transaction + if ( $this->mRecursive ) { + $this->queueRecursiveJobs(); + } - wfProfileOut( $fname ); + wfProfileOut( __METHOD__ ); } - function invalidateCategories( $cats ) { - $fname = 'LinksUpdate::invalidateCategories'; - if ( count( $cats ) ) { - $this->mDb->update( 'page', array( 'page_touched' => $this->mDb->timestamp() ), - array( - 'page_namespace' => NS_CATEGORY, - 'page_title IN (' . $this->mDb->makeList( array_keys( $cats ) ) . ')' - ), $fname + function queueRecursiveJobs() { + global $wgUpdateRowsPerJob; + wfProfileIn( __METHOD__ ); + + $cache = $this->mTitle->getBacklinkCache(); + $batches = $cache->partition( 'templatelinks', $wgUpdateRowsPerJob ); + if ( !$batches ) { + wfProfileOut( __METHOD__ ); + return; + } + $jobs = array(); + foreach ( $batches as $batch ) { + list( $start, $end ) = $batch; + $params = array( + 'start' => $start, + 'end' => $end, ); + $jobs[] = new RefreshLinksJob2( $this->mTitle, $params ); } + Job::batchInsert( $jobs ); + + wfProfileOut( __METHOD__ ); + } + + /** + * Invalidate the cache of a list of pages from a single namespace + * + * @param integer $namespace + * @param array $dbkeys + */ + function invalidatePages( $namespace, $dbkeys ) { + if ( !count( $dbkeys ) ) { + return; + } + + /** + * Determine which pages need to be updated + * This is necessary to prevent the job queue from smashing the DB with + * large numbers of concurrent invalidations of the same page + */ + $now = $this->mDb->timestamp(); + $ids = array(); + $res = $this->mDb->select( 'page', array( 'page_id' ), + array( + 'page_namespace' => $namespace, + 'page_title IN (' . $this->mDb->makeList( $dbkeys ) . ')', + 'page_touched < ' . $this->mDb->addQuotes( $now ) + ), __METHOD__ + ); + while ( $row = $this->mDb->fetchObject( $res ) ) { + $ids[] = $row->page_id; + } + if ( !count( $ids ) ) { + return; + } + + /** + * Do the update + * We still need the page_touched condition, in case the row has changed since + * the non-locking select above. + */ + $this->mDb->update( 'page', array( 'page_touched' => $now ), + array( + 'page_id IN (' . $this->mDb->makeList( $ids ) . ')', + 'page_touched < ' . $this->mDb->addQuotes( $now ) + ), __METHOD__ + ); + } + + function invalidateCategories( $cats ) { + $this->invalidatePages( NS_CATEGORY, array_keys( $cats ) ); + } + + /** + * Update all the appropriate counts in the category table. + * @param $added associative array of category name => sort key + * @param $deleted associative array of category name => sort key + */ + function updateCategoryCounts( $added, $deleted ) { + $a = new Article($this->mTitle); + $a->updateCategoryCounts( + array_keys( $added ), array_keys( $deleted ) + ); + } + + function invalidateImageDescriptions( $images ) { + $this->invalidatePages( NS_FILE, array_keys( $images ) ); } function dumbTableUpdate( $table, $insertions, $fromField ) { - $fname = 'LinksUpdate::dumbTableUpdate'; - $this->mDb->delete( $table, array( $fromField => $this->mId ), $fname ); + $this->mDb->delete( $table, array( $fromField => $this->mId ), __METHOD__ ); if ( count( $insertions ) ) { - # The link array was constructed without FOR UPDATE, so there may be collisions - # Ignoring for now, I'm not sure if that causes problems or not, but I'm fairly - # sure it's better than without IGNORE - $this->mDb->insert( $table, $insertions, $fname, array( 'IGNORE' ) ); + # The link array was constructed without FOR UPDATE, so there may + # be collisions. This may cause minor link table inconsistencies, + # which is better than crippling the site with lock contention. + $this->mDb->insert( $table, $insertions, __METHOD__, array( 'IGNORE' ) ); } } @@ -169,11 +305,15 @@ class LinksUpdate { /** * Update a table by doing a delete query then an insert query - * @access private + * @private */ function incrTableUpdate( $table, $prefix, $deletions, $insertions ) { - $fname = 'LinksUpdate::incrTableUpdate'; - $where = array( "{$prefix}_from" => $this->mId ); + if ( $table == 'page_props' ) { + $fromField = 'pp_page'; + } else { + $fromField = "{$prefix}_from"; + } + $where = array( $fromField => $this->mId ); if ( $table == 'pagelinks' || $table == 'templatelinks' ) { $clause = $this->makeWhereFrom2d( $deletions, $prefix ); if ( $clause ) { @@ -182,17 +322,24 @@ class LinksUpdate { $where = false; } } else { + if ( $table == 'langlinks' ) { + $toField = 'll_lang'; + } elseif ( $table == 'page_props' ) { + $toField = 'pp_propname'; + } else { + $toField = $prefix . '_to'; + } if ( count( $deletions ) ) { - $where[] = "{$prefix}_to IN (" . $this->mDb->makeList( array_keys( $deletions ) ) . ')'; + $where[] = "$toField IN (" . $this->mDb->makeList( array_keys( $deletions ) ) . ')'; } else { $where = false; } } if ( $where ) { - $this->mDb->delete( $table, $where, $fname ); + $this->mDb->delete( $table, $where, __METHOD__ ); } if ( count( $insertions ) ) { - $this->mDb->insert( $table, $insertions, $fname, 'IGNORE' ); + $this->mDb->insert( $table, $insertions, __METHOD__, 'IGNORE' ); } } @@ -200,7 +347,7 @@ class LinksUpdate { /** * Get an array of pagelinks insertions for passing to the DB * Skips the titles specified by the 2-D array $existing - * @access private + * @private */ function getLinkInsertions( $existing = array() ) { $arr = array(); @@ -221,7 +368,7 @@ class LinksUpdate { /** * Get an array of template insertions. Like getLinkInsertions() - * @access private + * @private */ function getTemplateInsertions( $existing = array() ) { $arr = array(); @@ -241,7 +388,7 @@ class LinksUpdate { /** * Get an array of image insertions * Skips the names specified in $existing - * @access private + * @private */ function getImageInsertions( $existing = array() ) { $arr = array(); @@ -257,7 +404,7 @@ class LinksUpdate { /** * Get an array of externallinks insertions. Skips the names specified in $existing - * @access private + * @private */ function getExternalInsertions( $existing = array() ) { $arr = array(); @@ -276,25 +423,64 @@ class LinksUpdate { * Get an array of category insertions * @param array $existing Array mapping existing category names to sort keys. If both * match a link in $this, the link will be omitted from the output - * @access private + * @private */ function getCategoryInsertions( $existing = array() ) { + global $wgContLang; $diffs = array_diff_assoc( $this->mCategories, $existing ); $arr = array(); foreach ( $diffs as $name => $sortkey ) { + $nt = Title::makeTitleSafe( NS_CATEGORY, $name ); + $wgContLang->findVariantLink( $name, $nt, true ); $arr[] = array( 'cl_from' => $this->mId, 'cl_to' => $name, - 'cl_sortkey' => $sortkey + 'cl_sortkey' => $sortkey, + 'cl_timestamp' => $this->mDb->timestamp() ); } return $arr; } + /** + * Get an array of interlanguage link insertions + * @param array $existing Array mapping existing language codes to titles + * @private + */ + function getInterlangInsertions( $existing = array() ) { + $diffs = array_diff_assoc( $this->mInterlangs, $existing ); + $arr = array(); + foreach( $diffs as $lang => $title ) { + $arr[] = array( + 'll_from' => $this->mId, + 'll_lang' => $lang, + 'll_title' => $title + ); + } + return $arr; + } + + /** + * Get an array of page property insertions + */ + function getPropertyInsertions( $existing = array() ) { + $diffs = array_diff_assoc( $this->mProperties, $existing ); + $arr = array(); + foreach ( $diffs as $name => $value ) { + $arr[] = array( + 'pp_page' => $this->mId, + 'pp_propname' => $name, + 'pp_value' => $value, + ); + } + return $arr; + } + + /** * Given an array of existing links, returns those links which are not in $this * and thus should be deleted. - * @access private + * @private */ function getLinkDeletions( $existing ) { $del = array(); @@ -311,7 +497,7 @@ class LinksUpdate { /** * Given an array of existing templates, returns those templates which are not in $this * and thus should be deleted. - * @access private + * @private */ function getTemplateDeletions( $existing ) { $del = array(); @@ -328,16 +514,16 @@ class LinksUpdate { /** * Given an array of existing images, returns those images which are not in $this * and thus should be deleted. - * @access private + * @private */ function getImageDeletions( $existing ) { return array_diff_key( $existing, $this->mImages ); } - /** - * Given an array of existing external links, returns those links which are not + /** + * Given an array of existing external links, returns those links which are not * in $this and thus should be deleted. - * @access private + * @private */ function getExternalDeletions( $existing ) { return array_diff_key( $existing, $this->mExternals ); @@ -346,20 +532,36 @@ class LinksUpdate { /** * Given an array of existing categories, returns those categories which are not in $this * and thus should be deleted. - * @access private + * @private */ function getCategoryDeletions( $existing ) { return array_diff_assoc( $existing, $this->mCategories ); } + /** + * Given an array of existing interlanguage links, returns those links which are not + * in $this and thus should be deleted. + * @private + */ + function getInterlangDeletions( $existing ) { + return array_diff_assoc( $existing, $this->mInterlangs ); + } + + /** + * Get array of properties which should be deleted. + * @private + */ + function getPropertyDeletions( $existing ) { + return array_diff_assoc( $existing, $this->mProperties ); + } + /** * Get an array of existing links, as a 2-D array - * @access private + * @private */ function getExistingLinks() { - $fname = 'LinksUpdate::getExistingLinks'; $res = $this->mDb->select( 'pagelinks', array( 'pl_namespace', 'pl_title' ), - array( 'pl_from' => $this->mId ), $fname, $this->mOptions ); + array( 'pl_from' => $this->mId ), __METHOD__, $this->mOptions ); $arr = array(); while ( $row = $this->mDb->fetchObject( $res ) ) { if ( !isset( $arr[$row->pl_namespace] ) ) { @@ -373,12 +575,11 @@ class LinksUpdate { /** * Get an array of existing templates, as a 2-D array - * @access private + * @private */ function getExistingTemplates() { - $fname = 'LinksUpdate::getExistingTemplates'; $res = $this->mDb->select( 'templatelinks', array( 'tl_namespace', 'tl_title' ), - array( 'tl_from' => $this->mId ), $fname, $this->mOptions ); + array( 'tl_from' => $this->mId ), __METHOD__, $this->mOptions ); $arr = array(); while ( $row = $this->mDb->fetchObject( $res ) ) { if ( !isset( $arr[$row->tl_namespace] ) ) { @@ -392,12 +593,11 @@ class LinksUpdate { /** * Get an array of existing images, image names in the keys - * @access private + * @private */ function getExistingImages() { - $fname = 'LinksUpdate::getExistingImages'; $res = $this->mDb->select( 'imagelinks', array( 'il_to' ), - array( 'il_from' => $this->mId ), $fname, $this->mOptions ); + array( 'il_from' => $this->mId ), __METHOD__, $this->mOptions ); $arr = array(); while ( $row = $this->mDb->fetchObject( $res ) ) { $arr[$row->il_to] = 1; @@ -408,12 +608,11 @@ class LinksUpdate { /** * Get an array of existing external links, URLs in the keys - * @access private + * @private */ function getExistingExternals() { - $fname = 'LinksUpdate::getExistingExternals'; $res = $this->mDb->select( 'externallinks', array( 'el_to' ), - array( 'el_from' => $this->mId ), $fname, $this->mOptions ); + array( 'el_from' => $this->mId ), __METHOD__, $this->mOptions ); $arr = array(); while ( $row = $this->mDb->fetchObject( $res ) ) { $arr[$row->el_to] = 1; @@ -424,12 +623,11 @@ class LinksUpdate { /** * Get an array of existing categories, with the name in the key and sort key in the value. - * @access private + * @private */ function getExistingCategories() { - $fname = 'LinksUpdate::getExistingCategories'; $res = $this->mDb->select( 'categorylinks', array( 'cl_to', 'cl_sortkey' ), - array( 'cl_from' => $this->mId ), $fname, $this->mOptions ); + array( 'cl_from' => $this->mId ), __METHOD__, $this->mOptions ); $arr = array(); while ( $row = $this->mDb->fetchObject( $res ) ) { $arr[$row->cl_to] = $row->cl_sortkey; @@ -437,5 +635,69 @@ class LinksUpdate { $this->mDb->freeResult( $res ); return $arr; } + + /** + * Get an array of existing interlanguage links, with the language code in the key and the + * title in the value. + * @private + */ + function getExistingInterlangs() { + $res = $this->mDb->select( 'langlinks', array( 'll_lang', 'll_title' ), + array( 'll_from' => $this->mId ), __METHOD__, $this->mOptions ); + $arr = array(); + while ( $row = $this->mDb->fetchObject( $res ) ) { + $arr[$row->ll_lang] = $row->ll_title; + } + return $arr; + } + + /** + * Get an array of existing categories, with the name in the key and sort key in the value. + * @private + */ + function getExistingProperties() { + $res = $this->mDb->select( 'page_props', array( 'pp_propname', 'pp_value' ), + array( 'pp_page' => $this->mId ), __METHOD__, $this->mOptions ); + $arr = array(); + while ( $row = $this->mDb->fetchObject( $res ) ) { + $arr[$row->pp_propname] = $row->pp_value; + } + $this->mDb->freeResult( $res ); + return $arr; + } + + + /** + * Return the title object of the page being updated + */ + function getTitle() { + return $this->mTitle; + } + + /** + * Return the list of images used as generated by the parser + */ + public function getImages() { + return $this->mImages; + } + + /** + * Invalidate any necessary link lists related to page property changes + */ + function invalidateProperties( $changed ) { + global $wgPagePropLinkInvalidations; + + foreach ( $changed as $name => $value ) { + if ( isset( $wgPagePropLinkInvalidations[$name] ) ) { + $inv = $wgPagePropLinkInvalidations[$name]; + if ( !is_array( $inv ) ) { + $inv = array( $inv ); + } + foreach ( $inv as $table ) { + $update = new HTMLCacheUpdate( $this->mTitle, $table ); + $update->doUpdate(); + } + } + } + } } -?>