* Unrelated to the ResourcesTest structure test as it isn't testing
any of core's resources.
* Moved to the libs/Minify repo in change Ia9018a966a5325d.
Change-Id: If87035777633635fe4ad0f01790e7084f56f65a0
This tests needs the database in various places:
SqlModuleDependencyStore, wfGetDB() for WikiModule::preloadTitleInfo,
and other things in WikiModule. Some of these might be avoided; for
instance, it could use KeyValueDependencyStore. Others can't really be
avoided as the relevant classes call wfGetDB (or equivalent methods in
MediaWikiServices) directly.
Given the importance of this test, I'd rather have it continue using the
database, so add it to the Database group.
Change-Id: I564a826e97967897a3538d74cdef6f3dbdd570f1
Going forward we do not want to use the targets system in the way
we traditionally used it. While updating existing usages will take
time it seems like it might be a good idea to prevent additional usages
being added. This list would gradually be updated as we update those
components.
Bug: T127268
Depends-On: Ibd7270abc5f223bc4f4814c896e7ef4d02d21386
Change-Id: I2e226c174be88caeb505738645cfa0582e39a197
To implement source maps, we want FileModule::getScript() to return
an array with path information, but that would break subclasses of
FileModule which concatenate to the return value of parent::getScript().
So allow scripts to be generated by a callback, eliminating the need for
concatenation in subclasses.
* Factor out most of the loop body of expandPackageFiles() into
expandFileInfo(). Add 'name' to the return value. Ensure 'filePath'
is always a FilePath object since we need that to safely return it
to ResourceLoader later. Document the return value.
* Make sure the base path is always set in FilePath objects returned by
expandFileInfo().
* Factor out the loop body of the final stage of file info expansion
into readFileInfo(). Retain filePath, do not unset it.
* Assert that $fileInfo['content'] is definitely set.
* Convert array_map() in getDefinitionSummary() to a loop.
* Migrate LanguageDataModule.
Bug: T47514
Change-Id: I97d61b5793159cea365740e0563f7b733e0f16de
Confirm that load.php can respond for all registered modules without
errors (except private modules, which can't we don't serve there).
Bug: T47514
Change-Id: I44655b2c05529ae719d71622f57bfed8d632550e
The UA sniffs that overrode the feature tests are no longer needed.
* MSIE 10: Fine, rejected by feature checks.
* UC Mini "Speed Mode": Redundant, the version that this sniff
matched is pre-ES6. Current versions of UC Mini don't appear to
support enabling "Speed Mode" on random websites nor does it offer
it for Wikipedia specifically.
Details at https://phabricator.wikimedia.org/T178356#8740573.
* Google Web Light: Redundant, shutdown as of 2022.
Any references or extensions that still reach the proxy, get
redirected to our online URLs
https://googleweblight.com/?lite_url=https://en.m.wikipedia.org/wiki/Bananahttps://phabricator.wikimedia.org/T152602https://en.wikipedia.org/wiki/Google_Web_Light
* MeeGo: Redundant, discontinued and presumed rejected.
Either way, unsupported.
* Opera Mini: Fine, rejected by checks.
Details at https://phabricator.wikimedia.org/T178356#8740573.
* Ovi Browser: Redundant, discontinued and presumed rejected.
Either way, unsupported.
* Google Glass: Improve UX (since 2013, T58008).
* NetFront: Redundant. Old versions are presumed rejected.
Current versions are Chromium-based and presumed fine.
The exclusion was not UX based, but due to jQuery explicitly not
supporting it in 2013. This is no longer the case, so we can let
the feature test lead the way here.
* PlayStation: Redundant, same story as NetFront.
The version that matched the sniff is presumed rejected.
Current versions probably fine, but even not, don't match
our sniff so are already enabled today.
Bug: T178356
Change-Id: Ib6263ce3ffd11af5e501de8857f3e48a248c6210
If not specified, going forward all modules which do not define
targets should be assumed to be mobile AND desktop. I think this is
a fair expectation given that most new features should be built with
mobile in mind. This also avoids the issue where features are
unexpectedly not shipped to mobile.
The work in T324723 has ensured that all modules that need to remain
desktop only have been marked as such. After the audit, we can be
reasonably confident that enabling the remaining modules is okay,
given these modules run in isolated parts of the UI (not on every page
view).
I'm assuming CI is catching all offenders here. If not, any impacted
modules would require a trivial update to explicitly define modules
as mobile only. This would allow us to see clearly which modules
are depending on the target system and strategize around updating
them.
Once this patch is merged we can focus on fixing the modules we've
identified as mobile or desktop targeted and finally dismantling the
targets system altogether.
Bug: T127268
Change-Id: Ia062ff2d8b8732b0d3498c1a614bbf6a3e3a7ddb
Usually it doesn't make much of a difference, but potentially can
when the array is big. The extra conversion also makes the code more
complicated, I would argue.
Change-Id: I91b9e82ca801dd64523557d616462b7ea7edc7b0
With PHPUnit 9, tests without @covers tags are considered risky and emit
warnings. Not having @covers is bad practice anyway, so use
@coversNothing instead to make the intention clear, and re-enable the
phpcs rule.
Also rewrite an assertion in ResourcesTest that was bothering me.
Bug: T243600
Change-Id: I6dd683f93b6b2faed5f107be2ca7860602277fbc
We cannot access MW services in a data provider, as MW will not have
bootstrapped by this point in plain PHPUnit. And we do not want to
bootstrap MW by the time a dataProvider is executing, because unit tests
should not need (nor be able to access) MW services.
Change-Id: I21a5993f91b8fbe3fa476a8244b8911cf81e1d5f
Move ResourceLoader classes to their own namespace. Strip the
"ResourceLoader" prefix from all except ResourceLoader itself.
Move the tests by analogy.
I used a namespace alias "RL" in some callers since RL\Module is less
ambiguous at the call site than just "Module".
I did not address DependencyStore which continues to have a non-standard
location and namespace.
Revert of a241d83e0a.
Bug: T308718
Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
This reverts commit e08ea8ccb9.
Reason for revert: Breaks Phan in extensions, and as far as I’m aware,
this change isn’t urgently needed for anything, so the simplest fix is
to revert it again for now. After PHP 7.4 it should be safer to try this
again (we hopefully won’t need the two “hack” classes by then).
Bug: T308443
Change-Id: Iff3318cbf97a67f821f78e60da62a583f63e389e
Move ResourceLoader classes to their own namespace. Strip the
"ResourceLoader" prefix from all except ResourceLoader and
ResourceLoaderContext.
Move the tests by analogy.
I used a namespace alias "RL" in some callers since RL\Module is less
ambiguous at the call site than just "Module".
I did not address DependencyStore which continues to have a non-standard
location and namespace.
Change-Id: I92998ae6a82e0b935c13e02a183e7c324fa410a3
In all cases where the client registry is used, the `noscript` module
is inapplicable. Additionally, make it illegal to have a module
include the 'noscript' module as a dependency.
Bug: T291735
Change-Id: I5846f2c19b003989b56b12628b385fd37049483b
- JavaScriptMinifier: The public interface for this internal class is
ResourceLoader::filter which, except for one caller (minify.php)
is indeed consistently used already, no other callers need replacing.
- CSSMin: Idem for minification, however, this class has some other
method as well, which have a number of internal users, but none
outside core. These have all been replaced in this commit.
- Remove pear/net_url2 as own dependency as this was only used by CSSMin
and recently added there. This is now a dependency of wikimedia/minify
instead. This was not part of the public API and no longer mentioned
here in composer.json or RELEASE-NOTES. (It remains pinned in
mediawiki-vendor, however.)
Bug: T273247
Depends-On: Iadff8c2112d5e53bd994ab4882006e8c644a2379
Depends-On: Id3c6dcc0b952a9efe34b3a9ed88b716101a51f87
Depends-On: I589be910f57289fd908b22db87241b0e52da60d4
Change-Id: I8d71fc64aeecdb31db218b02e361ae9bb0d19b48
The name change happened some time ago, and I think its
about time to start using the name name!
(Done with a find and replace)
My personal motivation for doing this is that I have started
trying out vscode as an IDE for mediawiki development, and
right now it doesn't appear to handle php aliases very well
or at all.
Change-Id: I412235d91ae26e4c1c6a62e0dbb7e7cf3c5ed4a6
ResourceLoaderFileModule::readStyleFile() both reads a style file from
disk and processes it through LESS, CSSJanus and CSSMin. Factor out the
processing part into ResourceLoaderFileModule::processStyle(), which
takes a string of unprocessed CSS/LESS rather than a file. This is
needed for future support for styles that don't come (directly) from a
file, but are generated through packageFiles.
Other changes:
- Use a hash of the source instead of the file name in the cache keys
for the LESS compilation output
- Also prefix the cache key with 'resourceloader' while we're changing
it anyway
- LESS compilation no longer adds the source file itself as a dependency
- Don't pass down $flip, just use $this>getFlip()
Change-Id: I86e880d06af724f0fbae93e042c85e0395771912
Avoid data providers for dynamic expansion over all JS, CSS
and images files known to MW and extensions (1000+), because
there's non-trivival overhead in PHPUnit for creating a separate
"test" wrapper and iteration for each.
Instead, make each group a single test with each of the checks
being an assertion within that.
Reduces test count of 'php phpunit.php structure/ResourcesTest.php'
(with MW core + Vector + Minerva installed) from 1039 tests to
37 tests, and reduces runtime from ~1.6min to ~40sec.
This is the same pattern I previously adapted in a5a672b1a5
for the biggest part test of this suite (dependency verification),
but I forgot to apply it to all the others.
Bug: T225730
Change-Id: Iae46b2d9ff8cd70aa6e64aaddf6ccf839591153b
New data provider and test method were added to ensure file paths are correct and the file is present.
Bug: T231430
Change-Id: Iadafe0df1104de8829ea87632d65b3008cf80d76
This isn't needed because the startup module validates this already.
The vast majority of modules are FileModule instances which can't be invalid,
because a separate test asserts that class already. This test existed for
validating the format of version hashes returned by a theoretical Module
sub class in an extension that (badly) overrides the getVersionHash method.
As of writing, no extension overrides that method. And more importantly,
the startup module already validates this at run-time, and logs a warning.
This commit turns that into an exception, which would get logged in a way
that Jenkins will fail the build if encountered.
This structure test, which computed the response for all registered modules,
previously took 3-5 seconds in CI.
Bug: T225730
Change-Id: Id2e37434b0ccd95dd2279f04e2230e9c06b09ccb
Being a raw module means that when it is requested from load.php with
"only=scripts" set, then the output is *not* wrapped in an
'mw.loader.implement' closure *and* there no 'mw.loader.state()' appendix.
Instead, it is served "raw".
Before 2018, the modules 'mediawiki' and 'jquery' were raw modules.
They were needed before the client could define 'mw.loader.implement', and
could never be valid dependencies. Module 'mediawiki' merged to 'startup',
and 'jquery' became a regular module (T192623). Based on the architecture
of modules being deliverable bundles, it doesn't make sense for there to
ever be raw modules again. Anything that 'startup' needs should be bundled
with it. Anything else is a regular module.
On top of that, we never actually needed this feature because specifying
the 'only=scripts' and 'raw=1' parameters does the same thing.
The only special bit about marking modules (not requests) as "raw" was that
it allowed the client to forget to specify "raw=1" and the server would
automatically omit the 'mw.loader.state()' appendix based on whether the
module is marked as raw. As of Ie4564ec8e26ad53f2, the two remaining use
cases for raw responses now specify the 'raw=1' request parameter, and we
can get rid of the "raw module" feature and all the complexity around it.
== Startup module
In the startup module there was an interesting use of isRaw() that has
little to do with the above. The "ATTENTION" warning there applies to the
startup module only, not raw modules in general. This is now fixed by
explicitly checking for StartupModule.
Above that warning, it talked about saving bytes, which was an optimisation
given that "raw" modules don't communicate with mw.loader, they also don't
need to be registered there because even if mw.loader would try to load
them, the server would never inform mw.loader about the module having
arrived. There are now no longer any such modules.
Bug: T201483
Change-Id: I8839036e7b2b76919b6cd3aa42ccfde4d1247899
Previously removed from includes/ already.
Also remove odd `@package`, which we never use.
And remove `@since` which doesn't make sense for test-only files.
Change-Id: Ib7265d39329ecadd5279b11820f77f54189b55d2
@covers does not make any sense for structure tests, which either
do not cover any PHP lines (they test things like configuration or
messages), or cover lines which cannot be determined at the time
of writing the tests (e.g. they cover all classes implementing a
certain interface). Requiring @coversNothing to be manually added
for all of them is a waste of developer time.
tests/phpunit/suite.xml has forceCoversAnnotation=true so removing
the annotations will not change test coverage, these files will
still be skipped.
Change-Id: I27cb58e92341b9b1a76f109701f5bc843adbaa9b
This issue came up in I8a49143, see
https://integration.wikimedia.org/ci/job/mediawiki-quibble-vendor-postgres-php70-docker/2453/console
The third parameter of assertArrayEquals() is called $ordered and is
meant to take the order of elements into account. Providing a string sets
this to true. The SQL query in ChangesListSpecialPageTest seems to behave
a bit random in Postgres and does not always return the elements in the
same order. This is fine. It's just the assertion that was to strict, by
accident.
I found a few more instances of the same issue with a regular expression.
In most cases I intentionally changed it to assertSame() because the order
of elements is actually guaranteed by the code, and needs to be (e.g.
mixing width and height of an image would be fatal).
Change-Id: Ice66cab873a7271d55809a486ce28cf637e43e33
Deprecated since MW 1.27.
Also update ResourcesTest to use TestingAccessWrapper instead of long-form
object reflection, and also apply it to its call for this method given
its meant to be private.
Change-Id: I9cc1af93730f632e4f8bf3a16d514a51ee73cb03
It also removes some code duplication which is nice.
This unlocks various future changes, including:
* Making the `$config` parameter mandatory for the ResourceLoader class
constructor, which currently falls back to global state.
This should be deprecated and removed.
* Making it possible to instantiate the ResourceLoader class
without all the default MW modules being registered from
global state. E.g. move MW module registration from main class
constructor to ServiceWiring, and remove the 'EmptyResourceLoader'
class hack from unit tests, and use regular 'new ResourceLoader'
instead.
* Making ResourceLoader a standalone library (some day),
e.g. allowing it to be instantiated from a basic PHP script,
in a way that is still useful and perhaps able to serve
(most) RL modules without MW itself.
Bug: T32956
Change-Id: I4939f296c705b268e9cf8de635e923a739410470
Embed the essential files to define mw.loader directly as part of
the startup module.
* This means the internal 'mediawiki' module no longer exists.
This is safe to remove because:
1) While registered server-side for loading from startup.js, a PHPUnit
structure test disallowed being specified as a dependency.
2) Anything that attempted to load it client-side failed because the
module was marked in the registry as 'raw', thereby excluding it
from the data sent to the client-side. As such, it was seen as an
unknown module that the client refused to fetch from the server.
* Deprecate getStartupModules() and getLegacyModules().
These are no longer needed. There are no known callers anywhere in
Wikimedia Git or elsewhere indexed by Codesearch, but easy enough
to leave as no-op for one release.
* Remove ResourceLoaderRawFileModule class.
No longer needed. Was created as a hack specifically for the 'mediawiki'
module so that it would not leak global variables in debug mode.
It has no usage anywhere in Wikimedia Git, nor elsewhere in Codesearch.
Remove without deprecation given this was meant to be a 'private' class.
* Introduce (private) getBaseModules(). Previously, this list only existed
locally in getStartupModulesUrl() by merging getStartupModules() and
getLegacyModules(). This value was factored out into its own method.
* Make getStartupModulesUrl() private and rename to getBaseModulesUrl().
It is only used internally to export the 'baseModulesUri' value.
Its name was already confusing before, but it would've been even more
confusing now given it doesn't even call getStartupModules() any more.
Bug: T192623
Change-Id: I14ba282d7b65e99ca54b7c2f77ba6e1adaddd11c
* Use getStartupModules() instead of hardcoding these two module names.
* Also collect any other module that is registered with raw=>true.
Change-Id: I5f75cc8f99b90f58d76b7e87f068f50f9f608ed0
This is a pure documentation change. It mostly removes empty lines from
comments (and entirely empty comments), as well as adds a few missing
documentation blocks and fixes a minor mistake. I hope it's ok to have
this in one patch. I can split it, please tell me.
Change-Id: I9668338602ac77b903ab6b02ff56bd52743c37c4
