aboutsummaryrefslogtreecommitdiffstats
path: root/includes/MediaWikiEntryPoint.php
Commit message (Collapse)AuthorAgeFilesLines
* Namespace all remaining files in includes/exceptionJames D. Forrester2025-03-251-1/+1
| | | | | | | | | In MediaWiki/Exception, to follow PSR-4 per plural vs. singular (this can be changed later if people really care). Also, move the couple of exceptions in here that were already namespaced in the MW-top-level into the new space. Bug: T353458 Change-Id: I12ed850ae99effb699a6d7ada173f54e72f0570e
* Namespace all remaining files in includes/jobqueueJames D. Forrester2025-03-201-2/+2
| | | | | Bug: T353458 Change-Id: I95690a312e356c45dbeed607d32fb0e4626690cf
* Remove 2-line PHPDocs that just repeat the types from the codethiemowmde2025-01-171-4/+0
| | | | | | | | | | | | | | | | Same as Ia294bf4 did for 1-line comments. This patch removes slightly more complex 2-line PHPDoc comments that don't add any new information to the code, but literally repeat what the code already says. They say "don't document the code, code the documentation", and we are doing this more and more. We just tend to forget to remove the obsolete comments. Note I'm also removing a line of text in a few cases when it's very short and literally says the same as the method name. Again, such comments add zero new information. Change-Id: I01535404bab458c6c47e48e5456403b7a64198ed
* Tracing: span statuses for root spansChris Danis2025-01-071-1/+6
| | | | | Bug: T340552 Change-Id: I18ffd0e36efbabe82313de685ca8c6c99341c6f3
* Stats: Move StatsFactory flush into emitBufferedStatsCole White2024-12-031-7/+11
| | | | | | | | | | | | This centralizes the action of emitting stats to a single static function to be used by web and maintenance jobs alike. * Rename emitBufferedStatsdData to emitBufferedStats * Move StatsFactory->flush() call into emitBufferedStats * Update call signature where used Bug: T380609 Change-Id: I63e8832a170bfaa0e14e977f0852c4ef49f232bb
* Add request-level OpenTelemetry instrumentationMáté Szabó2024-11-261-0/+5
| | | | | | | | | | | | | | | | | | | | Why: - Now that we have a basic OpenTelemetry library available in MediaWiki, we would like to start instrumenting code and propagating the trace context. What: - Start a root span in early setup code once the service container is available, so that subsequent setup logic (e.g. session initialization) can already be instrumented. - End the root span in restInPeace() after the main transaction round ended, so that the normal LBFactory shutdown process can be instrumented. - Have TracerSharedState hold the root span throughout. Bug: T340552 Change-Id: I0785204a6989787a492a08b78c40c6df91106f23
* Add namespace to remaining parts of Wikimedia\Mime and Wikimedia\StatsJames D. Forrester2024-09-271-1/+1
| | | | | Bug: T353458 Change-Id: If0137003ab625017d322d57870448a02569668c3
* Merge "entrypoint: Do not add new headers when headers were already sent."jenkins-bot2024-08-091-0/+1
|\
| * entrypoint: Do not add new headers when headers were already sent.Gergő Tisza2024-08-081-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Do not try to add any extra headers in MediaWikiEntryPoint::outputResponsePayload() when headers have already been sent, it would just result in outputting a warning and messing up the response body. This can happen when output is managed manually (OutputPage::disable() has been called), for example in the CentralAuth extension SpecialCentralAutoLogin::doFinalOutput() would otherwise break because it will output JS without any output buffering, and then outputResponsePayload() would try to send a header. Bug: T364116 Change-Id: Id55b76274393dec21ab1d26aff0c1e8a63c5b28d
* | Stats: Remove $wgSamplingStatsdClient, deprecate SamplingStatsdClient classTimo Tijhof2024-08-071-3/+2
|/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | == Motivation == * Reduce amount of configuration switches relating to MW Stats component, to reduce noise for MW sysadmins and devs. * Reduce amount of code in the legacy Statsd implementation. == Background == * commit e56f7b6c63 (2015, MW 1.26): Add statsd sampling. Refers to T106457 under T106450 ("hook usage counts"), which used it in https://gerrit.wikimedia.org/r/226640 which was reverted shortly after in https://gerrit.wikimedia.org/r/233045 due to too high overhead of stats code for something as hot as hooks, regardless of actual sampling/stats sending. * commit b203ec5fcd (2016, MW 1.29): Make statsd sampling configurable. Use case unknown, but it is not used today in wmf-config, and appears to have never been set in the past either: ``` operations-mediawiki-config (master) $ git log -p wmf-config/ | grep wgStatsdSamplingRates 24s ec=1$ ``` Looking for unmerged patches, I do find an abandoned patch for the Wikibase extension that attempted to use to instrument database load in 2020. It was abandoned in favour of dedicated DB tooling to inspect database queries, which offer more detail as well. https://gerrit.wikimedia.org/r/q/message:wgStatsdSamplingRates https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Wikibase/+/593335 == Future == In 10 years, we've not needed sampling anywhere even at our scale. StatsD absorbs high traffic very well, with controlled flushing to Graphite as backend. Prometheus exporters are similarly buffered close to the producer with controlled scraping intervals. This should make sampling unlikely to be needed. We did have one experiment where sampling was attempted, which led to the developments this commit removes. That instrumentation was removed because the code was far too frequently called and hot that it posed a latency problem, even before we factor in whether it sends the metric. The overhead of the stats code itself added too much overhead to Hook::run / HookContainer. The new StatsFactory service class does feature per-metric sampling, although this is similarly without use case at the moment: https://codesearch.wmcloud.org/search/?q=%3EsetSampleRate&excludeFiles=test New instrumentations should use StatsFactory, and if sampling were needed in the future, it could be experimented by calling setSampleRate on Metric objects (as part of the instrumentation, rather than in site config, unless the feature in question makes it configurable). That seems enough upfront investment for a theoretical future need. Change-Id: I5f68e48d6d2c880a43f83915234d3c9f32a2d1ef
* Fix use of stale backlink partition cacheTim Starling2024-06-281-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If a template is created, and then used on a page, and then the template is updated, all within an hour, then the page_touched of the page is never updated and the user will always see the old template contents. This is because htmlCacheUpdate jobs are fully suppressed for one hour following template creation, due to the WAN backlink partition cache. So: * Revert 4f3efbf4065a9, so that htmlCacheUpdate jobs always do something regardless of the state of the partition cache. * Factor out the job queueing parts of WikiPage::onArticleCreate, ::onArticleDelete and ::onArticleEdit. Instead of queueing a job unconditionally, check for the existence of backlinks in a post-send deferred update. If there are none, don't queue the job. * It's convenient to use BacklinkCache::hasLinks(), however, it suffered from the same stale cache problem as BacklinkCache::partition(). It's a short and fast query, and code review shows that none of the callers are particularly performance sensitive. So, do not use the WAN cache in BacklinkCache::hasLinks(). * Since hasLinks() and getNumLinks() no longer share a significant amount of code, separate them. Remove the $max parameter from getNumLinks(), which only existed to support hasLinks() and has no other usages in codesearch. * Log a debug message when entering the post-send request stage, so that it's easier to confirm that no additional pre-send queries are done. * Add a regression test, confirmed to previously fail. Bug: T368006 Change-Id: Id5c7af6d4fcdbeb6724a9036133742c5f76624df
* filerepo: extract AuthenticatedFileEntryPoint from img_auth.phpdaniel2024-05-161-0/+1
| | | | | | | | | The idea is that all entry points should share the code in the MediaWikiEntryPoint base class. This change just moves code from the file scope into a class, without any structural changes. Bug: T354216 Change-Id: Ie2e827d30a070bcc63bdce56891c3aa0a4dacddd
* Introduce entry point classes for media files.daniel2024-03-301-1/+2
| | | | | | | | | | | | | This moves a code out of file scope into classes to make it testable. The code is left in the same structure as it was before, global functions have been converted into methods on the new ThumbnailEntryPoint and Thumbnail404EntryPoint classes. This test introduces comprehensive phpunit tests covering all functional code paths in ThumbnailEntryPoint. This is intended to support refactoring of this code. Change-Id: I459abc7b11d0ab4ee682a863c9525a945048296f
* ResourceLoader: Follow-up creation of ResourceLoaderEntryPointTimo Tijhof2024-02-221-8/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Follows-up Iadea44b7867f48a2be6ccbf00c0e56911a5af74e. ResourceLoaderEntryPoint: * Rename from Resource to ResourceLoader, as it is the entry point for ResourceLoader. * Fix accidental use of ingroup on the file docblock instead of the class. https://phabricator.wikimedia.org/F42046908 * Fix broken "see" reference to "/load.php" on the ResourceEntryPoint class on doc.wikimedia.org. Files are indexed in Doxygen by full name without leading slash. * Remove load.php from docgroup, akin to index.php. tests: * Remove unrelated Rest namespace from test suite. * Remove redundant default values in newResourceLoader(), especially null values were confusing by showing an apparent intent to set a value without actually isplacing the default, given that isset/?? treat set null as not set. * Simplify test by using the built-in startup module instead. EntryPoint: * Remove comment in test about doPostOutputShutdown. It sounded like it was explaining that we use getCapturedOutput() instead of run() because the latter calls doPostOutputShutdown which we want to avoid. Except, we don't avoid it, since run() is called and does call doPostOutputShutdown. Unclear what it was explaining. * Improve docs around getCapturedOutput() in the base class, and make explicit that this seemingly unused complexity exists for the purpose of testing. Bug: T354216 Change-Id: I5139fda90a3d5ac7ea1ed917c8045d1a09961d78
* Merge "flushOutputBuffer: improve error handling"jenkins-bot2024-02-111-48/+99
|\
| * flushOutputBuffer: improve error handlingdaniel2024-02-091-48/+99
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This improves handling of edge cases around flushing buffers: - Distingush clearly between "headers sent" and "all output sent" states. - Warn when trying to flush output after output was already sent to the client. - Avoid looping indefinitely when failing to delete a buffer, to work around a quirk of PHP's zlib.compression buffer. Follow-up: I4ea116d60030da92be14f0283ce4dc4877c4ca13 Bug: T356578 Change-Id: Iaf34a48b89e969b8e26328f0d2512ad22ef48212
* | Namespace includes/contextJames D. Forrester2024-02-081-1/+1
|/ | | | | Bug: T353458 Change-Id: I4dbef138fd0110c14c70214282519189d70c94fb
* Don't assume no prior output in MWEntryPoint::flushOutputBuffer()Gergő Tisza2024-02-041-6/+1
| | | | | | | | | | This is not the method that actually flushes the output buffer under normal circumstances (that's done in execute(), within e.g. MediaWikiEntryPoint::outputResponsePayload() or ResourceLoader::respond()). Also fix a misleading phpdoc. Bug: T356578 Change-Id: I4ea116d60030da92be14f0283ce4dc4877c4ca13
* ResourceLoader: Factor out ResourceEntryPoint from load.phpdaniel2024-02-021-3/+5
| | | | | | | | The idea is that all entry points should share the code in the MediaWikiEntryPoint base class. This change by itself doesn't achieve much, it should be followed up by a change moving request handling logic from ResourceLoader into ResourceEntryPoint. Bug: T354216 Change-Id: Iadea44b7867f48a2be6ccbf00c0e56911a5af74e
* entrypoint: Isolate entry points from PHP runtime for testingdaniel2024-02-021-31/+388
| | | | | | | | | | | | 1) Introduce EntryPointEnvironment which wraps functions that interact with the PHP runtime, so they can be mocked for testing. 2) Allow server info fields to be overwritten in FauxRequest. 3) Make MediaWikiEntryPoint use WebResponse to set headers Bug: T354216 Change-Id: Ic21950c956de5d2b5a7dd66a1e2de58f807cfd9f
* MediaWikiEntryPoint: Inject MediaWiki services containerDerick Alangi2024-01-241-19/+71
| | | | | | | | | | | In order for sub-classes to benefit from this abstract class, inject the services container to lazily get all service dependencies needed within calling code. Define protected getters that can be overriden or called directly to get the service needed. Change-Id: I23a60907805dfb8fb7ce8160df63ababefcd62c3
* stats: add metric stats.statslib.bufferedCole White2024-01-151-3/+3
| | | | | | | | | | | | | | | | | | Records the number of samples of all metrics in the buffered statslib cache. Necessitates adding getSampleCount() to BaseMetric and wiring access through each metric type. Moved StatsFactory->flush() to execute before emitBufferedStatsdData(). This is so the metric generated by copyToStatsdAt() is sent with the rest of the statsd metrics. Updated tests. This metric will be used to track the migration progress towards statslib. Bug: T350638 Change-Id: I070f3e30a9b8cd0b0ab44f345207e6f25e602fbb
* MediaWikiEntryPoint: add metric stats.statsdclient.bufferedCole White2024-01-151-0/+1
| | | | | | | | | | Sends the number of samples in the buffered statsd data to graphite. This metric will be used to track the migration progress towards statslib. Bug: T350638 Change-Id: I12c42c51bd0488777474057c6c485743aa916ca8
* entrypoint: Restore doc grouping for MediaWikiEntryPointTimo Tijhof2024-01-051-0/+10
| | | | | | | | | | | | | | | Follows-up Ib70e4e67e4cb1. * This removed index.php from the "entrypoint" group, but did not add MediaWikiEntryPoint in its stead. * Fix broken `@see` reference, leading slash is not part of the identifier for the `@file` block index.php, only works without it. * Move the defgroup to a more natural place in the parent class' file. Bug: T354216 Change-Id: I1281e0934368884f4d1ffbbd3f5588b9a211356f
* Introduce MediaWikiEntryPoint and ActionEntryPointdaniel2024-01-031-0/+777
This extracts the MediaWikiEntryPoint base class and ActionEntryPoint class from the MediaWiki class. MediaWiki itself be deprecated. The intent is to create other subclasses of MediaWikiEntryPoint for the use by other entry points such as thumb.php or api.php. This will allow us to share code between entry points, and make these entry points testable by moving their implementation into a class. Bug: T354216 Change-Id: Ib70e4e67e4cb1b65ac218c095864fb6eb43d0929
generated by cgit v1.2.3 (git 2.25.1) at 2025-04-21 08:16:41 +0000