Before this commit, the reason set in the global $wgReadOnly was differently handled
on different special pages. While on most of them, like Special:Upload, the reason
is allowed to have HTML, which can be used in Wikitext, too, Special:CreateAccount
always outputted an escaped version of this reason.
Most special pages uses the ReadOnlyError exception to print a read-only error,
however, AuthManager uses Status objects to communicate between the backend and the
frontend. Therefore the same message and parameters were wrapped in a Status object
and, in the frontend, directly passed to the constructor of ErrorPageError. Unfortunately,
Status::getMessage() escapes the parameters of a message, which is the reason, why the
wiki is read-only. To bypass this restriction, AuthManager now creates a Message object
directly, does not escape the reason, and uses the resulting object to create a Status
object from.
Now the reason is not escaped on Special:CreateAccount anymore, like on most other
special pages.
The read-only message on the protection form is, also before this commit, not escaped and
already displayed correctly, as the read-only is checked in the constructor of the
protection form already and, if the Wiki is read only, handled as a permission error and
already displayed correctly. This commit fixes the behavior of WikiPage in case of it's used
somewhere else, subclassed or if the check in the frontend will be removed and the Status of
WikiPage will be used.
Bug: T157036
Change-Id: Idbfe556fcb90f8bda8fae9d728ca9dee5ea02f67
For some varargs a variable name is added with suffix ,... as seen for
many other varargs
Some @param are swapped, because there are in the wrong order
Enable Sniff MediaWiki.Commenting.FunctionComment.ParamNameNoMatch
Change-Id: I60fec6025bce824d5c67563ab7b65ad6cd628ad8
Guarded by the $wgUsePigLatinVariant variable, off by default.
Pig Latin is a language game where words in English are altered
according to the following rules:
* Words starting with a vowel have a '-way' suffix appended.
* Words starting with a consonant have the initial consonants (or 'qu'
group) moved to the end and an '-ay' suffix appended.
https://en.wikipedia.org/wiki/Pig_Latin
* Added 'en-x-piglatin' as a language name.
* Added 'en' to LanguageConverter::$languagesWithVariants.
* Added LanguageEn class and its corresponding EnConverter which
provides one-way translation from English to Pig Latin.
* Some minor internal changes in code that assumed that English
doesn't have a language class or converter.
Bug: T45547
Depends-On: I1d9691c784032669979f8109c9a5f65cbf4122c9
Change-Id: I7fa2d85d6364958c5138366e8b4504a2697a8731
Replaces \TestingAccessWrapper (defined in core) with
\Wikimedia\TestingAccessWrapper (defined in the composer package
wikimedia/testing-access-wrapper).
See https://gerrit.wikimedia.org/r/#/q/topic:librarize-testing-access-wrapper
for downstream patches.
The core version of the class is kept around for a while to avoid
circular dependency problems.
Bug: T163434
Change-Id: I52cc257e593da3d6c3b01a909e554a950225aec8
Introduce a service to represent wfReadOnly() and friends.
It's necessary to have two service instances, one for wfReadOnly() and
one for wfConfiguredReadOnlyReason(), to avoid a circular dependency,
since LoadBalancer needs the configured reason during construction, but
wfReadOnly() needs to query the currently active load balancer.
Not having a cache of the configuration makes it possible to dynamically
change the configuration. Ideally things would not change the
configuration, and I removed such instances in core, but to support
extensions, I added a test ensuring that the configuration can be changed.
Change-Id: I9bbee946c10742526d3423208efd68cb3cc5a7ee
The default will remain PHPUnit 4.x due to PHP 5.5 support.
But, we should allow developers to run tests with newer PHPUnit
versions which are noticably faster (especially for code coverage
reports).
* <https://github.com/sebastianbergmann/phpunit/wiki/Release-Announcement-for-PHPUnit-5.4.0>
PHPUnit 5 deprecates the getMock() shortcut for getMockBuilder()->getMock().
It instead introduces the shortcut createMock() which has better defaults
than getMockBuilder(). For example, it sets 'disableArgumentCloning' and
other things by default.
Going forward, code should either use getMockBuilder directly and configure
it using the setter methods (instead of the confusing variadic arguments
of getMock) or simply use the new minimalistic createMock method. This patch
backports the createMock method to MediaWikiTestCase so that we can start
using it.
Change-Id: I091c0289b21d2b1c876adba89529dc3e72b99af2
Creating users involves database access and the test
database isn't fully setup yet when data providers run.
In this particular test / data provider, all that is needed
is the user name so we can just set that w/o creating
the user in the data provider. The user gets created during
the test. (after the test database is setup)
also $id set in $id = $user->getId() was unused and was the
specific code that involved making a database connection.
Change-Id: I1685dfdd6cfc60cd2c59a4d9de95bcbbaba5e153
If anyone wants such a thing, they can make their own extension.
I asked stewards, and they said they don't use this.
See also T32636 / 9de2bfd1fe
Bug: T150930
Change-Id: I3ab5962dba668e5d628e55ad0c0feae471d82b5e
The `type` log event attribute is used in Wikimedia production logging
to categorize log events by source (e.g. 'mediawiki', 'restbase', etc).
Adding a `type` key to the logging context overwrites the default value
of 'mediawiki'. Rename the key to `throttle` in the context and the
message template.
Change-Id: Ic274159774e43a8749f83c850fff7897956cf377
When the local password provider is in login-only mode, it should
disable itself as soon as some other primary provider is enabled.
It's impossible to tell whether that is the case though, so err in
the safer direction.
Change-Id: Ie77a9cc6d8f06aa52a893e753d9971e30b0f55e5
It looks like there is something missing after the last statement
Also remove some other empty lines at begin of functions, ifs or loops
while at these files
Change-Id: Ib00b5cfd31ca4dcd0c32ce33754d3c80bae70641
Logstash merges the log context into the main metadata (where
'message' is the log message) and ends up overwriting the message.
Bug: T145133
Change-Id: I27f221b0f1f7203e93d1b92119dc584ba8526f5b
No idea what's causing T146498, but maybe this cleanup will fix it.
Note we can't clean up every use because we can't create test users from
data providers.
Bug: T146498
Change-Id: I5c7ea7f49ed42cb2ed36460afcf6aec30e9359de
This adds support for multiple count-per-period limits and arbitrary
period durations in the AuthManager account creation throttle in the
wiki settings. The $wgAccountCreationThrottle config variable becomes
an array like $wgPasswordAttemptThrottle.
Bug: T146290
Change-Id: Iea182a92a1199b0ce7103ab9ae24f1c87b01985c
This commit changes the way how HTMLForm handles a Status object
when executed from a request. It now handles, beside the errors,
also the warnings of a Status object and prints them out, wrapped
in a warning box.
The LoginSignupPage uses this feature to show informative warnings
actually as warnings and not as more disturbing error messages.
Error messages should be reserved for errors and only for erros. An
AuthenticationProvider, which returns an UI AuthenticationResponse
can choose, if the given message is an error or a warning message.
This commit also addds a new function to Status, which allows a
developer to split the object into two new Status objects, where one only
contains the errors and the other only the warnings of the origin
Status object (splitByErrorType). StatusValue also has a new function,
splitByErrorType(), to support this.
Bug: T139179
Change-Id: I9a27911613e62b5c4cb86bea40696cb37c4f49c2
AuthManager tries to check whether the user already exists if
User::addToDatabase fails in autocreation, but since the same DB row
was already checked a few lines earlier and this method is typically
wrapped in an implicit transaction, it will just re-read the same
snapshot and not do anything useful. addToDatabase already has
a check for that so let's rely on that instead.
Bug: T145131
Change-Id: I94a5e8b851dcf994f5f9e773edf4e9153a4a3535
This can allow AuthenticationRequests to flag certain fields as
sensitive, so e.g. the API can insist they be in the POST body rather
than in the query string.
Change-Id: I7b12aa4cd8f5a570f0df7213c0f9084b5a4d4de7
AuthManager::getAuthenticationRequests() changes
AuthenticationRequest::$required from REQUIRED to PRIMARY_REQUIRED
if the request is from a primary; it made an exception when
all primary providers returned a given request. That exception is
not particularly useful (AuthenticationRequest::mergeFieldInfo()
used to rely on it to determine which fields are required, but
since I9d33bd2 that's not really needed), and knowing which request
is from a primary is useful for other means.
This changes required field semantics in a corner case: when a
primary provider returns two required requests, the previous
behavior was to assume that they are both required; the new one
is to treat them as alternatives (as if they were returned by
two different providers). So when all primary providers return
request X, and one of them returns Y in addition, the fields of X
will not be marked required, while previously that would have been
the case.
Instead of overcomplicating the interface for something that is
unlikely to come up in any real use case, add a new requirement
to PrimaryAuthenticationProvider that it should not return
multiple required requests.
Bug: T141471
Change-Id: I1c1f44d4d6b66f77c876e3459fb97f03483db744
Simply testing shouldn't call AbortNewAccount, we only want to do that
when the account is actually being created.
Change-Id: Icb3d1ce63a2691aa232b4564ed88fee6d50d7ab7
I726d79de added validation of the incoming requests to
ConfirmLinkSecondaryAuthenticationProvider, but forgot to fill in the
username and action fields before doing so meaning the confirmation
would probably fail.
Change-Id: I75cb65c24538a1e60e3b47a3b9ecfbd3a79ce8bf
They were coming out as null instead, which screws up when requests are
changing their fields based on the action.
Change-Id: Ic8caf57ebad35c3eb17d45f9d96c6de5b559a83a
* ApiQueryAuthManagerInfo will differentiate between preserved linking
data and a preserved createRequest.
* ApiQueryAuthManagerInfo will indicate the preserved username, if any,
because the client will have to pass that back to action=createaccount.
* ApiClientLogin won't tell about the confusing
CreateFromLoginAuthenticationRequest returned on RESTART responses.
* Explain how 'preservestate' works in ApiAMCreateAccount's auto-doc.
* ConfirmLinkSecondaryAuthenticationProvider will filter out requests
that can no longer be used (i.e. if it was for linking the account
that got used for creation).
* All the complicated code in AuthManager::beginAccountCreation() was
trying to deal with allowing the client to pass only the
CreateFromLoginAuthenticationRequest. That was dumb, removed it.
* Added methods to CreateFromLoginAuthenticationRequest to indicate its
status with respect to different kinds of preserved state.
* Increase accuracy of the AuthenticationResponse::$createRequest doc.
Change-Id: I726d79de18e739d6e60c1eea51453433c21ba207
Instead of only flagging fields which are required by a request
needed by all primairy providers, it should be enough if all
requests needed by some primary provider require that field.
Also make CreationReasonAuthenticationRequest non-required so that
the list of required form fields is more in sync with that of
pre-AuthManager code.
Bug: T85853
Change-Id: I9d33bd22295758cc532a260b1848616b41d94f12
The unit tests spend nearly half of their run time resetting the user table for
each test. But the majority of tests do not depend on the user table having the
exact value that the setup code resets it to, and do not need to modify the
user objects they require to run.
Fix that by providing an API for tests to get User objects, and to indicate
whether the User object will be subject to destructive modification or not.
This allows User objects to be reused across multiple unit tests.
Change-Id: I17ef1f519759c5e7796c259282afe730ef722e96
Rewrite authentication-related special pages to use AuthManager.
All the changes mentioned below only take effect when
$wgDisableAuthManager is false.
LoginForm is rewritten to use HTMLForm and split into UserLogin
and CreateAccount; ChangePassword and PasswordReset are rewritten;
ChangeEmail and Preferences are updated. Four new special pages
are added to handle the new capabilities of AuthManager (linked
accounts, secondary authentication providers): LinkAccounts,
UnlinkAccounts, ChangeCredentials, RemoveCredentials.
The old form-based hooks (ChangePasswordForm, UserCreateForm,
UserLoginForm) are deprecated. A new, more generic hook is
available to alter the forms (AuthChangeFormFields);
form changes that involve new fields should be done via
$wgAuthManagerConfig.
UserLoginComplete is limited to web-based login; for more
generic functionality UserLoggedIn can be used instead.
Hooks that assume password-based login (PrefsPasswordAudit,
AbortChangePassword) are removed; the first functionality
is replaced by ChangeAuthenticationDataAudit, the second is
handled by AuthManager. LoginPasswordResetMessage is removed,
the functionality can be recreated via authentication providers.
There are several smaller backwards incompatible changes:
* Adding fields to the login/signup forms by manipulating the
template via the extraInput/extrafields parameters is not
supported anymore. Depending on the authn configuration the
login/signup process might be multistep and it would be
complicated to ensure that extensions can access the data
at the right moment. Instead, you can create an
AuthenticationProvider which can define its own fields and
process them when the authentication is over.
(There is B/C support for a transitional period that works with
the default login form, but might break with configurations that
require multiple steps or redirects.)
* Removed cookie redirect check. This was added in 2003 in 9ead07fe9
for the benefit of bots, but with MediaWiki having an API these days
there is little reason to keep it. Same for the wpSkipCookieCheck
flag (added in 2008 in 29c73e8265).
* Instead of embedding a password field on sensitive special pages
such as ChangeEmail, such pages rely on AuthManager for elevated
security (which typically involves requiring the user to log in again
unless their last login was more than a few minutes ago).
Accordingly, wgRequirePasswordforEmailChange is removed.
* Special:ChangePassword requires login now.
* Special:ResetPassword now sends a separate email to each user when called
with a shared email address.
* the Reason field had a message with 'prefsectiontip' class
which was sorta broken but used in extensions for formatting.
HTMLForm does not support that, so this commit turns it into a help message
which will break formatting. See https://gerrit.wikimedia.org/r/#/c/231884
Bug: T110277
Change-Id: I8b52ec8ddf494f23941807638f149f15b5e46b0c
Depends-On: If4e0dfb6ee6674f0dace80a01850e2d0cbbdb47a
This implements the AuthManager class and its needed interfaces and
subclasses, and integrates them into the backend portion of MediaWiki.
Integration with frontend portions of MediaWiki (e.g. ApiLogin,
Special:Login) is left for a followup.
Bug: T91699
Bug: T71589
Bug: T111299
Co-Authored-By: Gergő Tisza <gtisza@wikimedia.org>
Change-Id: If89d24838e326fe25fe867d02181eebcfbb0e196