wiki.techinc.nl/includes/MediaWiki.php
Timo Tijhof 11e3172c03 exception: Let MediaWiki.php control final output for ErrorPageError
The same way it does already for non-error output. This makes
it so that doPreOutputCommit() consistently happens between
the staging of output and the actual sending of output.

It is still allowed for code to bypass this, such as for fatal
errors and for handlers that disable OutputPage (like Special:Export).

But for cases where we do want to perform doPreOutputCommit(), it
should be run consistently between staging and sending so that it
can make appropiate decisions based on the current state of
OutputPage.

Previously, the state of OutputPage seen by doPreOutputCommit()
would be the broken/incomplete output of a seemingly succesful
(possibly cacheable) user action, which would then, after
doPreOutputCommit() runs, be completely replaced by $e->report()/
$out->showErrorPage().

This is a prerequisite for being able to reliably send cookie-block
cookies on error pages (next patch).

Bug: T233594
Change-Id: Iaeaf5e55a5868e6be534ddda73f3b56b9d6ef8f0
2019-09-26 07:56:14 +00:00

1125 lines
39 KiB
PHP

<?php
/**
* Helper class for the index.php entry point.
*
* 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
*/
use MediaWiki\Logger\LoggerFactory;
use Psr\Log\LoggerInterface;
use MediaWiki\MediaWikiServices;
use Wikimedia\Rdbms\ILBFactory;
use Wikimedia\Rdbms\ChronologyProtector;
use Wikimedia\Rdbms\DBConnectionError;
use Liuggio\StatsdClient\Sender\SocketSender;
/**
* The MediaWiki class is the helper class for the index.php entry point.
*/
class MediaWiki {
/**
* @var IContextSource
*/
private $context;
/**
* @var Config
*/
private $config;
/**
* @var string Cache what action this request is
*/
private $action;
/**
* @param IContextSource|null $context
*/
public function __construct( IContextSource $context = null ) {
if ( !$context ) {
$context = RequestContext::getMain();
}
$this->context = $context;
$this->config = $context->getConfig();
}
/**
* Parse the request to get the Title object
*
* @throws MalformedTitleException If a title has been provided by the user, but is invalid.
* @return Title Title object to be $wgTitle
*/
private function parseTitle() {
$request = $this->context->getRequest();
$curid = $request->getInt( 'curid' );
$title = $request->getVal( 'title' );
$action = $request->getVal( 'action' );
if ( $request->getCheck( 'search' ) ) {
// Compatibility with old search URLs which didn't use Special:Search
// Just check for presence here, so blank requests still
// show the search page when using ugly URLs (T10054).
$ret = SpecialPage::getTitleFor( 'Search' );
} elseif ( $curid ) {
// URLs like this are generated by RC, because rc_title isn't always accurate
$ret = Title::newFromID( $curid );
} else {
$ret = Title::newFromURL( $title );
// Alias NS_MEDIA page URLs to NS_FILE...we only use NS_MEDIA
// in wikitext links to tell Parser to make a direct file link
if ( !is_null( $ret ) && $ret->getNamespace() == NS_MEDIA ) {
$ret = Title::makeTitle( NS_FILE, $ret->getDBkey() );
}
$contLang = MediaWikiServices::getInstance()->getContentLanguage();
// Check variant links so that interwiki links don't have to worry
// about the possible different language variants
if (
$contLang->hasVariants() && !is_null( $ret ) && $ret->getArticleID() == 0
) {
$contLang->findVariantLink( $title, $ret );
}
}
// If title is not provided, always allow oldid and diff to set the title.
// If title is provided, allow oldid and diff to override the title, unless
// we are talking about a special page which might use these parameters for
// other purposes.
if ( $ret === null || !$ret->isSpecialPage() ) {
// We can have urls with just ?diff=,?oldid= or even just ?diff=
$oldid = $request->getInt( 'oldid' );
$oldid = $oldid ?: $request->getInt( 'diff' );
// Allow oldid to override a changed or missing title
if ( $oldid ) {
$rev = Revision::newFromId( $oldid );
$ret = $rev ? $rev->getTitle() : $ret;
}
}
// Use the main page as default title if nothing else has been provided
if ( $ret === null
&& strval( $title ) === ''
&& !$request->getCheck( 'curid' )
&& $action !== 'delete'
) {
$ret = Title::newMainPage();
}
if ( $ret === null || ( $ret->getDBkey() == '' && !$ret->isExternal() ) ) {
// If we get here, we definitely don't have a valid title; throw an exception.
// Try to get detailed invalid title exception first, fall back to MalformedTitleException.
Title::newFromTextThrow( $title );
throw new MalformedTitleException( 'badtitletext', $title );
}
return $ret;
}
/**
* Get the Title object that we'll be acting on, as specified in the WebRequest
* @return Title
*/
public function getTitle() {
if ( !$this->context->hasTitle() ) {
try {
$this->context->setTitle( $this->parseTitle() );
} catch ( MalformedTitleException $ex ) {
$this->context->setTitle( SpecialPage::getTitleFor( 'Badtitle' ) );
}
}
return $this->context->getTitle();
}
/**
* Returns the name of the action that will be executed.
*
* @return string Action
*/
public function getAction() {
if ( $this->action === null ) {
$this->action = Action::getActionName( $this->context );
}
return $this->action;
}
/**
* Performs the request.
* - bad titles
* - read restriction
* - local interwiki redirects
* - redirect loop
* - special pages
* - normal pages
*
* @throws MWException|PermissionsError|BadTitleError|HttpError
* @return void
*/
private function performRequest() {
global $wgTitle;
$request = $this->context->getRequest();
$requestTitle = $title = $this->context->getTitle();
$output = $this->context->getOutput();
$user = $this->context->getUser();
if ( $request->getVal( 'printable' ) === 'yes' ) {
$output->setPrintable();
}
$unused = null; // To pass it by reference
Hooks::run( 'BeforeInitialize', [ &$title, &$unused, &$output, &$user, $request, $this ] );
// Invalid titles. T23776: The interwikis must redirect even if the page name is empty.
if ( is_null( $title ) || ( $title->getDBkey() == '' && !$title->isExternal() )
|| $title->isSpecial( 'Badtitle' )
) {
$this->context->setTitle( SpecialPage::getTitleFor( 'Badtitle' ) );
try {
$this->parseTitle();
} catch ( MalformedTitleException $ex ) {
throw new BadTitleError( $ex );
}
throw new BadTitleError();
}
// Check user's permissions to read this page.
// We have to check here to catch special pages etc.
// We will check again in Article::view().
$permErrors = $title->isSpecial( 'RunJobs' )
? [] // relies on HMAC key signature alone
: $title->getUserPermissionsErrors( 'read', $user );
if ( count( $permErrors ) ) {
// T34276: allowing the skin to generate output with $wgTitle or
// $this->context->title set to the input title would allow anonymous users to
// determine whether a page exists, potentially leaking private data. In fact, the
// curid and oldid request parameters would allow page titles to be enumerated even
// when they are not guessable. So we reset the title to Special:Badtitle before the
// permissions error is displayed.
// The skin mostly uses $this->context->getTitle() these days, but some extensions
// still use $wgTitle.
$badTitle = SpecialPage::getTitleFor( 'Badtitle' );
$this->context->setTitle( $badTitle );
$wgTitle = $badTitle;
throw new PermissionsError( 'read', $permErrors );
}
// Interwiki redirects
if ( $title->isExternal() ) {
$rdfrom = $request->getVal( 'rdfrom' );
if ( $rdfrom ) {
$url = $title->getFullURL( [ 'rdfrom' => $rdfrom ] );
} else {
$query = $request->getValues();
unset( $query['title'] );
$url = $title->getFullURL( $query );
}
// Check for a redirect loop
if ( !preg_match( '/^' . preg_quote( $this->config->get( 'Server' ), '/' ) . '/', $url )
&& $title->isLocal()
) {
// 301 so google et al report the target as the actual url.
$output->redirect( $url, 301 );
} else {
$this->context->setTitle( SpecialPage::getTitleFor( 'Badtitle' ) );
try {
$this->parseTitle();
} catch ( MalformedTitleException $ex ) {
throw new BadTitleError( $ex );
}
throw new BadTitleError();
}
// Handle any other redirects.
// Redirect loops, titleless URL, $wgUsePathInfo URLs, and URLs with a variant
} elseif ( !$this->tryNormaliseRedirect( $title ) ) {
// Prevent information leak via Special:MyPage et al (T109724)
$spFactory = MediaWikiServices::getInstance()->getSpecialPageFactory();
if ( $title->isSpecialPage() ) {
$specialPage = $spFactory->getPage( $title->getDBkey() );
if ( $specialPage instanceof RedirectSpecialPage ) {
$specialPage->setContext( $this->context );
if ( $this->config->get( 'HideIdentifiableRedirects' )
&& $specialPage->personallyIdentifiableTarget()
) {
list( , $subpage ) = $spFactory->resolveAlias( $title->getDBkey() );
$target = $specialPage->getRedirect( $subpage );
// Target can also be true. We let that case fall through to normal processing.
if ( $target instanceof Title ) {
if ( $target->isExternal() ) {
// Handle interwiki redirects
$target = SpecialPage::getTitleFor(
'GoToInterwiki',
'force/' . $target->getPrefixedDBkey()
);
}
$query = $specialPage->getRedirectQuery( $subpage ) ?: [];
$request = new DerivativeRequest( $this->context->getRequest(), $query );
$request->setRequestURL( $this->context->getRequest()->getRequestURL() );
$this->context->setRequest( $request );
// Do not varnish cache these. May vary even for anons
$this->context->getOutput()->lowerCdnMaxage( 0 );
$this->context->setTitle( $target );
$wgTitle = $target;
// Reset action type cache. (Special pages have only view)
$this->action = null;
$title = $target;
$output->addJsConfigVars( [
'wgInternalRedirectTargetUrl' => $target->getFullURL( $query ),
] );
$output->addModules( 'mediawiki.action.view.redirect' );
}
}
}
}
// Special pages ($title may have changed since if statement above)
if ( $title->isSpecialPage() ) {
// Actions that need to be made when we have a special pages
$spFactory->executePath( $title, $this->context );
} else {
// ...otherwise treat it as an article view. The article
// may still be a wikipage redirect to another article or URL.
$article = $this->initializeArticle();
if ( is_object( $article ) ) {
$this->performAction( $article, $requestTitle );
} elseif ( is_string( $article ) ) {
$output->redirect( $article );
} else {
throw new MWException( "Shouldn't happen: MediaWiki::initializeArticle()"
. " returned neither an object nor a URL" );
}
}
}
}
/**
* Handle redirects for uncanonical title requests.
*
* Handles:
* - Redirect loops.
* - No title in URL.
* - $wgUsePathInfo URLs.
* - URLs with a variant.
* - Other non-standard URLs (as long as they have no extra query parameters).
*
* Behaviour:
* - Normalise title values:
* /wiki/Foo%20Bar -> /wiki/Foo_Bar
* - Normalise empty title:
* /wiki/ -> /wiki/Main
* /w/index.php?title= -> /wiki/Main
* - Don't redirect anything with query parameters other than 'title' or 'action=view'.
*
* @param Title $title
* @return bool True if a redirect was set.
* @throws HttpError
*/
private function tryNormaliseRedirect( Title $title ) {
$request = $this->context->getRequest();
$output = $this->context->getOutput();
if ( $request->getVal( 'action', 'view' ) != 'view'
|| $request->wasPosted()
|| ( $request->getCheck( 'title' )
&& $title->getPrefixedDBkey() == $request->getVal( 'title' ) )
|| count( $request->getValueNames( [ 'action', 'title' ] ) )
|| !Hooks::run( 'TestCanonicalRedirect', [ $request, $title, $output ] )
) {
return false;
}
if ( $this->config->get( 'MainPageIsDomainRoot' ) && $request->getRequestURL() === '/' ) {
return false;
}
if ( $title->isSpecialPage() ) {
list( $name, $subpage ) = MediaWikiServices::getInstance()->getSpecialPageFactory()->
resolveAlias( $title->getDBkey() );
if ( $name ) {
$title = SpecialPage::getTitleFor( $name, $subpage );
}
}
// Redirect to canonical url, make it a 301 to allow caching
$targetUrl = wfExpandUrl( $title->getFullURL(), PROTO_CURRENT );
if ( $targetUrl == $request->getFullRequestURL() ) {
$message = "Redirect loop detected!\n\n" .
"This means the wiki got confused about what page was " .
"requested; this sometimes happens when moving a wiki " .
"to a new server or changing the server configuration.\n\n";
if ( $this->config->get( 'UsePathInfo' ) ) {
$message .= "The wiki is trying to interpret the page " .
"title from the URL path portion (PATH_INFO), which " .
"sometimes fails depending on the web server. Try " .
"setting \"\$wgUsePathInfo = false;\" in your " .
"LocalSettings.php, or check that \$wgArticlePath " .
"is correct.";
} else {
$message .= "Your web server was detected as possibly not " .
"supporting URL path components (PATH_INFO) correctly; " .
"check your LocalSettings.php for a customized " .
"\$wgArticlePath setting and/or toggle \$wgUsePathInfo " .
"to true.";
}
throw new HttpError( 500, $message );
}
$output->setCdnMaxage( 1200 );
$output->redirect( $targetUrl, '301' );
return true;
}
/**
* Initialize the main Article object for "standard" actions (view, etc)
* Create an Article object for the page, following redirects if needed.
*
* @return Article|string An Article, or a string to redirect to another URL
*/
private function initializeArticle() {
$title = $this->context->getTitle();
if ( $this->context->canUseWikiPage() ) {
// Try to use request context wiki page, as there
// is already data from db saved in per process
// cache there from this->getAction() call.
$page = $this->context->getWikiPage();
} else {
// This case should not happen, but just in case.
// @TODO: remove this or use an exception
$page = WikiPage::factory( $title );
$this->context->setWikiPage( $page );
wfWarn( "RequestContext::canUseWikiPage() returned false" );
}
// Make GUI wrapper for the WikiPage
$article = Article::newFromWikiPage( $page, $this->context );
// Skip some unnecessary code if the content model doesn't support redirects
if ( !ContentHandler::getForTitle( $title )->supportsRedirects() ) {
return $article;
}
$request = $this->context->getRequest();
// Namespace might change when using redirects
// Check for redirects ...
$action = $request->getVal( 'action', 'view' );
$file = ( $page instanceof WikiFilePage ) ? $page->getFile() : null;
if ( ( $action == 'view' || $action == 'render' ) // ... for actions that show content
&& !$request->getVal( 'oldid' ) // ... and are not old revisions
&& !$request->getVal( 'diff' ) // ... and not when showing diff
&& $request->getVal( 'redirect' ) != 'no' // ... unless explicitly told not to
// ... and the article is not a non-redirect image page with associated file
&& !( is_object( $file ) && $file->exists() && !$file->getRedirected() )
) {
// Give extensions a change to ignore/handle redirects as needed
$ignoreRedirect = $target = false;
Hooks::run( 'InitializeArticleMaybeRedirect',
[ &$title, &$request, &$ignoreRedirect, &$target, &$article ] );
$page = $article->getPage(); // reflect any hook changes
// Follow redirects only for... redirects.
// If $target is set, then a hook wanted to redirect.
if ( !$ignoreRedirect && ( $target || $page->isRedirect() ) ) {
// Is the target already set by an extension?
$target = $target ?: $page->followRedirect();
if ( is_string( $target ) && !$this->config->get( 'DisableHardRedirects' ) ) {
// we'll need to redirect
return $target;
}
if ( is_object( $target ) ) {
// Rewrite environment to redirected article
$rpage = WikiPage::factory( $target );
$rpage->loadPageData();
if ( $rpage->exists() || ( is_object( $file ) && !$file->isLocal() ) ) {
$rarticle = Article::newFromWikiPage( $rpage, $this->context );
$rarticle->setRedirectedFrom( $title );
$article = $rarticle;
$this->context->setTitle( $target );
$this->context->setWikiPage( $article->getPage() );
}
}
} else {
// Article may have been changed by hook
$this->context->setTitle( $article->getTitle() );
$this->context->setWikiPage( $article->getPage() );
}
}
return $article;
}
/**
* Perform one of the "standard" actions
*
* @param Page $page
* @param Title $requestTitle The original title, before any redirects were applied
*/
private function performAction( Page $page, Title $requestTitle ) {
$request = $this->context->getRequest();
$output = $this->context->getOutput();
$title = $this->context->getTitle();
$user = $this->context->getUser();
if ( !Hooks::run( 'MediaWikiPerformAction',
[ $output, $page, $title, $user, $request, $this ] )
) {
return;
}
$act = $this->getAction();
$action = Action::factory( $act, $page, $this->context );
if ( $action instanceof Action ) {
// Narrow DB query expectations for this HTTP request
$trxLimits = $this->config->get( 'TrxProfilerLimits' );
$trxProfiler = Profiler::instance()->getTransactionProfiler();
if ( $request->wasPosted() && !$action->doesWrites() ) {
$trxProfiler->setExpectations( $trxLimits['POST-nonwrite'], __METHOD__ );
$request->markAsSafeRequest();
}
# Let CDN cache things if we can purge them.
if ( $this->config->get( 'UseCdn' ) &&
in_array(
// Use PROTO_INTERNAL because that's what getCdnUrls() uses
wfExpandUrl( $request->getRequestURL(), PROTO_INTERNAL ),
$requestTitle->getCdnUrls()
)
) {
$output->setCdnMaxage( $this->config->get( 'CdnMaxAge' ) );
}
$action->show();
return;
}
// If we've not found out which action it is by now, it's unknown
$output->setStatusCode( 404 );
$output->showErrorPage( 'nosuchaction', 'nosuchactiontext' );
}
/**
* Run the current MediaWiki instance; index.php just calls this
*/
public function run() {
try {
$this->setDBProfilingAgent();
try {
$this->main();
} catch ( ErrorPageError $e ) {
$out = $this->context->getOutput();
// TODO: Should ErrorPageError::report accept a OutputPage parameter?
$e->report( ErrorPageError::STAGE_OUTPUT );
// T64091: while exceptions are convenient to bubble up GUI errors,
// they are not internal application faults. As with normal requests, this
// should commit, print the output, do deferred updates, jobs, and profiling.
$this->doPreOutputCommit();
$out->output(); // display the GUI error
}
} catch ( Exception $e ) {
$context = $this->context;
$action = $context->getRequest()->getVal( 'action', 'view' );
if (
$e instanceof DBConnectionError &&
$context->hasTitle() &&
$context->getTitle()->canExist() &&
in_array( $action, [ 'view', 'history' ], true ) &&
HTMLFileCache::useFileCache( $this->context, HTMLFileCache::MODE_OUTAGE )
) {
// Try to use any (even stale) file during outages...
$cache = new HTMLFileCache( $context->getTitle(), $action );
if ( $cache->isCached() ) {
$cache->loadFromFileCache( $context, HTMLFileCache::MODE_OUTAGE );
print MWExceptionRenderer::getHTML( $e );
exit;
}
}
MWExceptionHandler::handleException( $e );
} catch ( Error $e ) {
// Type errors and such: at least handle it now and clean up the LBFactory state
MWExceptionHandler::handleException( $e );
}
$this->doPostOutputShutdown( 'normal' );
}
private function setDBProfilingAgent() {
$services = MediaWikiServices::getInstance();
// Add a comment for easy SHOW PROCESSLIST interpretation
$name = $this->context->getUser()->getName();
$services->getDBLoadBalancerFactory()->setAgentName(
mb_strlen( $name ) > 15 ? mb_substr( $name, 0, 15 ) . '...' : $name
);
}
/**
* @see MediaWiki::preOutputCommit()
* @param callable|null $postCommitWork [default: null]
* @since 1.26
*/
public function doPreOutputCommit( callable $postCommitWork = null ) {
self::preOutputCommit( $this->context, $postCommitWork );
}
/**
* This function commits all DB and session changes as needed *before* the
* client can receive a response (in case DB commit fails) and thus also before
* the response can trigger a subsequent related request by the client
*
* If there is a significant amount of content to flush, it can be done in $postCommitWork
*
* @param IContextSource $context
* @param callable|null $postCommitWork [default: null]
* @since 1.27
*/
public static function preOutputCommit(
IContextSource $context, callable $postCommitWork = null
) {
$config = $context->getConfig();
$request = $context->getRequest();
$output = $context->getOutput();
$lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
// Try to make sure that all RDBMs, session, and other storage updates complete
ignore_user_abort( true );
// Commit all RDBMs changes from the main transaction round
$lbFactory->commitMasterChanges(
__METHOD__,
// Abort if any transaction was too big
[ 'maxWriteDuration' => $config->get( 'MaxUserDBWriteDuration' ) ]
);
wfDebug( __METHOD__ . ': primary transaction round committed' );
// Run updates that need to block the client or affect output (this is the last chance)
DeferredUpdates::doUpdates( 'run', DeferredUpdates::PRESEND );
wfDebug( __METHOD__ . ': pre-send deferred updates completed' );
// Persist the session to avoid race conditions on subsequent requests by the client
$request->getSession()->save(); // T214471
wfDebug( __METHOD__ . ': session changes committed' );
// Figure out whether to wait for DB replication now or to use some method that assures
// that subsequent requests by the client will use the DB replication positions written
// during the shutdown() call below; the later requires working around replication lag
// of the store containing DB replication positions (e.g. dynomite, mcrouter).
list( $flags, $strategy ) = self::getChronProtStrategy( $lbFactory, $output );
// Record ChronologyProtector positions for DBs affected in this request at this point
$cpIndex = null;
$cpClientId = null;
$lbFactory->shutdown( $flags, $postCommitWork, $cpIndex, $cpClientId );
wfDebug( __METHOD__ . ': LBFactory shutdown completed' );
$allowHeaders = !( $output->isDisabled() || headers_sent() );
if ( $cpIndex > 0 ) {
if ( $allowHeaders ) {
$now = time();
$expires = $now + ChronologyProtector::POSITION_COOKIE_TTL;
$options = [ 'prefix' => '' ];
$value = $lbFactory::makeCookieValueFromCPIndex( $cpIndex, $now, $cpClientId );
$request->response()->setCookie( 'cpPosIndex', $value, $expires, $options );
}
if ( $strategy === 'cookie+url' ) {
if ( $output->getRedirect() ) { // sanity
$safeUrl = $lbFactory->appendShutdownCPIndexAsQuery(
$output->getRedirect(),
$cpIndex
);
$output->redirect( $safeUrl );
} else {
$e = new LogicException( "No redirect; cannot append cpPosIndex parameter." );
MWExceptionHandler::logException( $e );
}
}
}
if ( $allowHeaders ) {
// Set a cookie to tell all CDN edge nodes to "stick" the user to the DC that
// handles this POST request (e.g. the "master" data center). Also have the user
// briefly bypass CDN so ChronologyProtector works for cacheable URLs.
if ( $request->wasPosted() && $lbFactory->hasOrMadeRecentMasterChanges() ) {
$expires = time() + $config->get( 'DataCenterUpdateStickTTL' );
$options = [ 'prefix' => '' ];
$request->response()->setCookie( 'UseDC', 'master', $expires, $options );
$request->response()->setCookie( 'UseCDNCache', 'false', $expires, $options );
}
// Avoid letting a few seconds of replica DB lag cause a month of stale data.
// This logic is also intimately related to the value of $wgCdnReboundPurgeDelay.
if ( $lbFactory->laggedReplicaUsed() ) {
$maxAge = $config->get( 'CdnMaxageLagged' );
$output->lowerCdnMaxage( $maxAge );
$request->response()->header( "X-Database-Lagged: true" );
wfDebugLog( 'replication',
"Lagged DB used; CDN cache TTL limited to $maxAge seconds" );
}
// Avoid long-term cache pollution due to message cache rebuild timeouts (T133069)
if ( MessageCache::singleton()->isDisabled() ) {
$maxAge = $config->get( 'CdnMaxageSubstitute' );
$output->lowerCdnMaxage( $maxAge );
$request->response()->header( "X-Response-Substitute: true" );
}
}
}
/**
* @param ILBFactory $lbFactory
* @param OutputPage $output
* @return array
*/
private static function getChronProtStrategy( ILBFactory $lbFactory, OutputPage $output ) {
// Should the client return, their request should observe the new ChronologyProtector
// DB positions. This request might be on a foreign wiki domain, so synchronously update
// the DB positions in all datacenters to be safe. If this output is not a redirect,
// then OutputPage::output() will be relatively slow, meaning that running it in
// $postCommitWork should help mask the latency of those updates.
$flags = $lbFactory::SHUTDOWN_CHRONPROT_SYNC;
$strategy = 'cookie+sync';
$allowHeaders = !( $output->isDisabled() || headers_sent() );
if ( $output->getRedirect() && $lbFactory->hasOrMadeRecentMasterChanges( INF ) ) {
// OutputPage::output() will be fast, so $postCommitWork is useless for masking
// the latency of synchronously updating the DB positions in all datacenters.
// Try to make use of the time the client spends following redirects instead.
$domainDistance = self::getUrlDomainDistance( $output->getRedirect() );
if ( $domainDistance === 'local' && $allowHeaders ) {
$flags = $lbFactory::SHUTDOWN_CHRONPROT_ASYNC;
$strategy = 'cookie'; // use same-domain cookie and keep the URL uncluttered
} elseif ( $domainDistance === 'remote' ) {
$flags = $lbFactory::SHUTDOWN_CHRONPROT_ASYNC;
$strategy = 'cookie+url'; // cross-domain cookie might not work
}
}
return [ $flags, $strategy ];
}
/**
* @param string $url
* @return string Either "local", "remote" if in the farm, "external" otherwise
*/
private static function getUrlDomainDistance( $url ) {
$clusterWiki = WikiMap::getWikiFromUrl( $url );
if ( WikiMap::isCurrentWikiId( $clusterWiki ) ) {
return 'local'; // the current wiki
}
if ( $clusterWiki !== false ) {
return 'remote'; // another wiki in this cluster/farm
}
return 'external';
}
/**
* This function does work that can be done *after* the
* user gets the HTTP response so they don't block on it
*
* This manages deferred updates, job insertion,
* final commit, and the logging of profiling data
*
* @param string $mode Use 'fast' to always skip job running
* @since 1.26
*/
public function doPostOutputShutdown( $mode = 'normal' ) {
// Record backend request timing
$timing = $this->context->getTiming();
$timing->mark( 'requestShutdown' );
// Perform the last synchronous operations...
try {
// Show visible profiling data if enabled (which cannot be post-send)
Profiler::instance()->logDataPageOutputOnly();
} catch ( Exception $e ) {
// An error may already have been shown in run(), so just log it to be safe
MWExceptionHandler::logException( $e );
}
// Disable WebResponse setters for post-send processing (T191537).
WebResponse::disableForPostSend();
$blocksHttpClient = true;
// Defer everything else if possible...
$callback = function () use ( $mode, &$blocksHttpClient ) {
try {
$this->restInPeace( $mode, $blocksHttpClient );
} catch ( Exception $e ) {
// If this is post-send, then displaying errors can cause broken HTML
MWExceptionHandler::rollbackMasterChangesAndLog( $e );
}
};
if ( function_exists( 'register_postsend_function' ) ) {
// https://github.com/facebook/hhvm/issues/1230
register_postsend_function( $callback );
/** @noinspection PhpUnusedLocalVariableInspection */
$blocksHttpClient = false;
} else {
if ( function_exists( 'fastcgi_finish_request' ) ) {
fastcgi_finish_request();
/** @noinspection PhpUnusedLocalVariableInspection */
$blocksHttpClient = false;
} else {
// Either all DB and deferred updates should happen or none.
// The latter should not be cancelled due to client disconnect.
ignore_user_abort( true );
}
$callback();
}
}
private function main() {
global $wgTitle;
$output = $this->context->getOutput();
$request = $this->context->getRequest();
// Send Ajax requests to the Ajax dispatcher.
if ( $request->getVal( 'action' ) === 'ajax' ) {
// Set a dummy title, because $wgTitle == null might break things
$title = Title::makeTitle( NS_SPECIAL, 'Badtitle/performing an AJAX call in '
. __METHOD__
);
$this->context->setTitle( $title );
$wgTitle = $title;
$dispatcher = new AjaxDispatcher( $this->config );
$dispatcher->performAction( $this->context->getUser() );
return;
}
// Get title from request parameters,
// is set on the fly by parseTitle the first time.
$title = $this->getTitle();
$action = $this->getAction();
$wgTitle = $title;
// Set DB query expectations for this HTTP request
$trxLimits = $this->config->get( 'TrxProfilerLimits' );
$trxProfiler = Profiler::instance()->getTransactionProfiler();
$trxProfiler->setLogger( LoggerFactory::getInstance( 'DBPerformance' ) );
if ( $request->hasSafeMethod() ) {
$trxProfiler->setExpectations( $trxLimits['GET'], __METHOD__ );
} else {
$trxProfiler->setExpectations( $trxLimits['POST'], __METHOD__ );
}
// If the user has forceHTTPS set to true, or if the user
// is in a group requiring HTTPS, or if they have the HTTPS
// preference set, redirect them to HTTPS.
// Note: Do this after $wgTitle is setup, otherwise the hooks run from
// isLoggedIn() will do all sorts of weird stuff.
if (
$request->getProtocol() == 'http' &&
// switch to HTTPS only when supported by the server
preg_match( '#^https://#', wfExpandUrl( $request->getRequestURL(), PROTO_HTTPS ) ) &&
(
$request->getSession()->shouldForceHTTPS() ||
// Check the cookie manually, for paranoia
$request->getCookie( 'forceHTTPS', '' ) ||
// check for prefixed version that was used for a time in older MW versions
$request->getCookie( 'forceHTTPS' ) ||
// Avoid checking the user and groups unless it's enabled.
(
$this->context->getUser()->isLoggedIn()
&& $this->context->getUser()->requiresHTTPS()
)
)
) {
$oldUrl = $request->getFullRequestURL();
$redirUrl = preg_replace( '#^http://#', 'https://', $oldUrl );
// ATTENTION: This hook is likely to be removed soon due to overall design of the system.
if ( Hooks::run( 'BeforeHttpsRedirect', [ $this->context, &$redirUrl ] ) ) {
if ( $request->wasPosted() ) {
// This is weird and we'd hope it almost never happens. This
// means that a POST came in via HTTP and policy requires us
// redirecting to HTTPS. It's likely such a request is going
// to fail due to post data being lost, but let's try anyway
// and just log the instance.
// @todo FIXME: See if we could issue a 307 or 308 here, need
// to see how clients (automated & browser) behave when we do
wfDebugLog( 'RedirectedPosts', "Redirected from HTTP to HTTPS: $oldUrl" );
}
// Setup dummy Title, otherwise OutputPage::redirect will fail
$title = Title::newFromText( 'REDIR', NS_MAIN );
$this->context->setTitle( $title );
// Since we only do this redir to change proto, always send a vary header
$output->addVaryHeader( 'X-Forwarded-Proto' );
$output->redirect( $redirUrl );
$output->output();
return;
}
}
if ( $title->canExist() && HTMLFileCache::useFileCache( $this->context ) ) {
// Try low-level file cache hit
$cache = new HTMLFileCache( $title, $action );
if ( $cache->isCacheGood( /* Assume up to date */ ) ) {
// Check incoming headers to see if client has this cached
$timestamp = $cache->cacheTimestamp();
if ( !$output->checkLastModified( $timestamp ) ) {
$cache->loadFromFileCache( $this->context );
}
// Do any stats increment/watchlist stuff, assuming user is viewing the
// latest revision (which should always be the case for file cache)
$this->context->getWikiPage()->doViewUpdates( $this->context->getUser() );
// Tell OutputPage that output is taken care of
$output->disable();
return;
}
}
// Actually do the work of the request and build up any output
$this->performRequest();
// GUI-ify and stash the page output in MediaWiki::doPreOutputCommit() while
// ChronologyProtector synchronizes DB positions or replicas across all datacenters.
$buffer = null;
$outputWork = function () use ( $output, &$buffer ) {
if ( $buffer === null ) {
$buffer = $output->output( true );
}
return $buffer;
};
// Now commit any transactions, so that unreported errors after
// output() don't roll back the whole DB transaction and so that
// we avoid having both success and error text in the response
$this->doPreOutputCommit( $outputWork );
// Now send the actual output
print $outputWork();
}
/**
* Ends this task peacefully
* @param string $mode Use 'fast' to always skip job running
* @param bool $blocksHttpClient Whether this blocks an HTTP response to a client
*/
public function restInPeace( $mode = 'fast', $blocksHttpClient = true ) {
$lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
// Assure deferred updates are not in the main transaction
$lbFactory->commitMasterChanges( __METHOD__ );
// Loosen DB query expectations since the HTTP client is unblocked
$trxProfiler = Profiler::instance()->getTransactionProfiler();
$trxProfiler->redefineExpectations(
$this->context->getRequest()->hasSafeMethod()
? $this->config->get( 'TrxProfilerLimits' )['PostSend-GET']
: $this->config->get( 'TrxProfilerLimits' )['PostSend-POST'],
__METHOD__
);
// Do any deferred jobs; preferring to run them now if a client will not wait on them
DeferredUpdates::doUpdates( $blocksHttpClient ? 'enqueue' : 'run' );
// Now that everything specific to this request is done,
// try to occasionally run jobs (if enabled) from the queues
if ( $mode === 'normal' ) {
$this->triggerJobs();
}
// Log profiling data, e.g. in the database or UDP
wfLogProfilingData();
// Commit and close up!
$lbFactory->commitMasterChanges( __METHOD__ );
$lbFactory->shutdown( $lbFactory::SHUTDOWN_NO_CHRONPROT );
wfDebug( "Request ended normally\n" );
}
/**
* Send out any buffered statsd data according to sampling rules
*
* @param IBufferingStatsdDataFactory $stats
* @param Config $config
* @throws ConfigException
* @since 1.31
*/
public static function emitBufferedStatsdData(
IBufferingStatsdDataFactory $stats, Config $config
) {
if ( $config->get( 'StatsdServer' ) && $stats->hasData() ) {
try {
$statsdServer = explode( ':', $config->get( 'StatsdServer' ), 2 );
$statsdHost = $statsdServer[0];
$statsdPort = $statsdServer[1] ?? 8125;
$statsdSender = new SocketSender( $statsdHost, $statsdPort );
$statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
$statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
$statsdClient->send( $stats->getData() );
$stats->clearData(); // empty buffer for the next round
} catch ( Exception $ex ) {
MWExceptionHandler::logException( $ex );
}
}
}
/**
* Potentially open a socket and sent an HTTP request back to the server
* to run a specified number of jobs. This registers a callback to cleanup
* the socket once it's done.
*/
public function triggerJobs() {
$jobRunRate = $this->config->get( 'JobRunRate' );
if ( $this->getTitle()->isSpecial( 'RunJobs' ) ) {
return; // recursion guard
} elseif ( $jobRunRate <= 0 || wfReadOnly() ) {
return;
}
if ( $jobRunRate < 1 ) {
$max = mt_getrandmax();
if ( mt_rand( 0, $max ) > $max * $jobRunRate ) {
return; // the higher the job run rate, the less likely we return here
}
$n = 1;
} else {
$n = intval( $jobRunRate );
}
$logger = LoggerFactory::getInstance( 'runJobs' );
try {
if ( $this->config->get( 'RunJobsAsync' ) ) {
// Send an HTTP request to the job RPC entry point if possible
$invokedWithSuccess = $this->triggerAsyncJobs( $n, $logger );
if ( !$invokedWithSuccess ) {
// Fall back to blocking on running the job(s)
$logger->warning( "Jobs switched to blocking; Special:RunJobs disabled" );
$this->triggerSyncJobs( $n, $logger );
}
} else {
$this->triggerSyncJobs( $n, $logger );
}
} catch ( JobQueueError $e ) {
// Do not make the site unavailable (T88312)
MWExceptionHandler::logException( $e );
}
}
/**
* @param int $n Number of jobs to try to run
* @param LoggerInterface $runJobsLogger
*/
private function triggerSyncJobs( $n, LoggerInterface $runJobsLogger ) {
$trxProfiler = Profiler::instance()->getTransactionProfiler();
$old = $trxProfiler->setSilenced( true );
try {
$runner = new JobRunner( $runJobsLogger );
$runner->run( [ 'maxJobs' => $n ] );
} finally {
$trxProfiler->setSilenced( $old );
}
}
/**
* @param int $n Number of jobs to try to run
* @param LoggerInterface $runJobsLogger
* @return bool Success
*/
private function triggerAsyncJobs( $n, LoggerInterface $runJobsLogger ) {
// Do not send request if there are probably no jobs
$group = JobQueueGroup::singleton();
if ( !$group->queuesHaveJobs( JobQueueGroup::TYPE_DEFAULT ) ) {
return true;
}
$query = [ 'title' => 'Special:RunJobs',
'tasks' => 'jobs', 'maxjobs' => $n, 'sigexpiry' => time() + 5 ];
$query['signature'] = SpecialRunJobs::getQuerySignature(
$query, $this->config->get( 'SecretKey' ) );
$errno = $errstr = null;
$info = wfParseUrl( $this->config->get( 'CanonicalServer' ) );
$host = $info ? $info['host'] : null;
$port = 80;
if ( isset( $info['scheme'] ) && $info['scheme'] == 'https' ) {
$host = "tls://" . $host;
$port = 443;
}
if ( isset( $info['port'] ) ) {
$port = $info['port'];
}
Wikimedia\suppressWarnings();
$sock = $host ? fsockopen(
$host,
$port,
$errno,
$errstr,
// If it takes more than 100ms to connect to ourselves there is a problem...
0.100
) : false;
Wikimedia\restoreWarnings();
$invokedWithSuccess = true;
if ( $sock ) {
$special = MediaWikiServices::getInstance()->getSpecialPageFactory()->
getPage( 'RunJobs' );
$url = $special->getPageTitle()->getCanonicalURL( $query );
$req = (
"POST $url HTTP/1.1\r\n" .
"Host: {$info['host']}\r\n" .
"Connection: Close\r\n" .
"Content-Length: 0\r\n\r\n"
);
$runJobsLogger->info( "Running $n job(s) via '$url'" );
// Send a cron API request to be performed in the background.
// Give up if this takes too long to send (which should be rare).
stream_set_timeout( $sock, 2 );
$bytes = fwrite( $sock, $req );
if ( $bytes !== strlen( $req ) ) {
$invokedWithSuccess = false;
$runJobsLogger->error( "Failed to start cron API (socket write error)" );
} else {
// Do not wait for the response (the script should handle client aborts).
// Make sure that we don't close before that script reaches ignore_user_abort().
$start = microtime( true );
$status = fgets( $sock );
$sec = microtime( true ) - $start;
if ( !preg_match( '#^HTTP/\d\.\d 202 #', $status ) ) {
$invokedWithSuccess = false;
$runJobsLogger->error( "Failed to start cron API: received '$status' ($sec)" );
}
}
fclose( $sock );
} else {
$invokedWithSuccess = false;
$runJobsLogger->error( "Failed to start cron API (socket error $errno): $errstr" );
}
return $invokedWithSuccess;
}
}