When this was originally written, the plan was to read both the old and
new fields during the transition period, while stopping writes to them
midway through. It turns out that the WHERE conditions to do read-both
correctly are generally not handled well by the database and working
around that would require a lot of complicated code (see what's being
removed from ApiQueryUserContribs here, for example).
We can simplify things greatly by instead having it write both fields
during the transition period, reading from the old for the first part
and the new for the second part, as is being done for MCR.
Bug: T204669
Change-Id: I4764c1c7883dc1003cb12729455c8107319f70b1
Depends-On: I845f6ae462f2539ebd35cbb5f2ca8b5714e2c1fb
Depends-On: I88b31b977543fdbdf69f8c1158e77e448df94e11
This introduces PageUpdater to replace WikiPage::doEditContent,
and DerivedPageDataUpdater, to replace WikiPage::doEditUpdates
and WikiPage::prepareContentForEdit.
See docs/pageupdater.txt for a description of their
functionality.
MCR migration notes:
* The interface of PageUpdater is expected to
remain mostly stable after this patch. Code that has been using
WikiPage::doEditContent can be confidently migrated to using the
new mechanism for revision creation.
* This patch keeps the code inside PageUpdater largely aligned
with the old code in WikiPage, to make review easier to to avoid
mistakes. It is intended to be refactored further, moving
application logic into stateless services.
* DerivedPageDataUpdate is intended as a stepping stone for further
refactoring. Its behavior is designed to be compatible with
callback code that currently relies on
WikiPage::prepareContentForEdit. Much of the code that currently
lives in DerivedPageDataUpdate should be factored out into
services, all behavior relevant to calling code should be exposed
via narrow interfaces.
Bug: T174038
Bug: T196653
Change-Id: If610c68f4912e89af616cdcac1d35a1be3946afa
Find: /isset\(\s*([^()]+?)\s*\)\s*\?\s*\1\s*:\s*/
Replace with: '\1 ?? '
(Everywhere except includes/PHPVersionCheck.php)
(Then, manually fix some line length and indentation issues)
Then manually reviewed the replacements for cases where confusing
operator precedence would result in incorrect results
(fixing those in I478db046a1cc162c6767003ce45c9b56270f3372).
Change-Id: I33b421c8cb11cdd4ce896488c9ff5313f03a38cf
It is a boolean parameter that is confusing for three reasons:
* It's a boolean parameter, given parameters are unnamed in PHP,
these are always poor UX for call sites.
* It's negated ("noudp"). save(true) means no feeds events,
save(false) [default] means sending events to feeds.
* To overcome this problem, typical use was to pass a free-form
string that self-documents the intended behaviour,
e.g. `save('pleasenoudp')`, any string casts to true.
Fix this by moving the booleans to constants and use those
instead. For compatiblity, keep the negation internally,
although it's hidden from regular usage. Also document
the string hack, deprecate it, and update callers.
Change-Id: Ia57c86b38bf50cb4ec580f42a6b1ca798fcf781a
These actions mark edits as patrolled automatically, so the
correct rc_patrolled value to use is PRC_AUTOPATROLLED rather than
PRC_PATROLLED. The code for log entries checks the autopatrol right, and
the code for category entries has a comment that says "just like log
entries".
Bug: T190408
Bug: T184791
Change-Id: Id994af8cd3833a862a09389431256aba35c2e6d9
Per the RFC, it will now become the default and only behaviour
to not log autpatrol actions. The information is already
recorded via the rc_patrolled field.
Bug: T184485
Change-Id: I98ae895a2b4cde4bb945f1df23be4a070b0bf9c4
Storing the user name or IP in every row in large tables like revision
and logging takes up space and makes operations on these tables slower.
This patch begins the process of moving those into one "actor" table
which other tables can reference with a single integer field.
A subsequent patch will remove the old columns.
Bug: T167246
Depends-On: I9293fd6e0f958d87e52965de925046f1bb8f8a50
Change-Id: I8d825eb02c69cc66d90bd41325133fd3f99f0226
This allows CommentStore to be added to MediaWikiServices
without the need of an aditional Factory.
This change includes a compatability layer to allow the behaviour
from 1.30 to continue to be used while deprecated.
CommentStore::newKey has been deprecated.
Keys are now passed into the public methods of CommentStore
where needed.
The following CommentStore methods have had their signatures changed
to introduced a $key parameter, but when used in conjunction with
CommentStore::newKey behaviour will remain unchanged:
* CommentStore::getFields
* CommentStore::getJoin
* CommentStore::getComment
* CommentStore::getCommentLegacy
* CommentStore::insert
* CommentStore::insertWithTemplate
Change-Id: I3abb62a5cfb0dcd456da9f4eb35583476ae41cfb
Several classes have a "selectFields()" static method to tell callers
which fields to select from the database. With the recent comment table
change and the upcoming actor table change, this pattern has become too
simplistic as a SELECT will need to join several tables to be able to
retrieve all the needed fields.
Thus, we deprecate the selectFields() methods in favor of getQueryInfo()
methods that return tables and join conditions in addition to the
fields.
Change-Id: Idcfd15568489d9f03a7ba4460e96610d33bc4089
CategoryMembershipChange is great, but it doesn't have a
machine-readable way to see if the category has been added or removed
from the page. We could parse the comments, but that would rely on
reading the comments and matching the message against the translated
strings.
This patch adds a couple of parameters and notes whether the change
was an addition or removal in rc_params.
Also, puts the evaluation of the AbortEmailNotification hook before
the check of rc_type in the hopes that the extension CatWatch can send
out emails at the right time.
Bug: T175052
Change-Id: I3e2ac0cc148b9d618e2f8b7853b14bff827ee443
Since the caller doesn't (and shouldn't) know whether CommentStore is
using the old or the new schema, it should leave truncation of comments
to CommentStore.
Change-Id: I92954c922514271d774518d6a6c28a01f33c88c2
And auto-fix all errors.
The `<exclude-pattern>` stanzas are now included in the default ruleset
and don't need to be repeated.
Change-Id: I928af549dc88ac2c6cb82058f64c7c7f3111598a
Follows-up 39a6e3dc4d. Class-based feeds are always given their parameters
by RCFeed::factory. However because the old getEngine() method insists
on creating its own object, the constructor parameters were not given.
Add it as optional parameter and pass it through there.
This is backwards-compatible still because before the 39a6e3dc4d refactor,
an RCFeedEngine also was not given information about any formatter and it
was the callers responsibility to format the line before calling send().
CentralAuth still uses it this way and that works fine. The core-caller
that expected the construction parameters since 39a6e3dc4d is hereby fixed.
The test couldn't catch this because it constructed the class instance there,
since PHPUnit does not support a mock class that is instantiated by foreign
code, and the parameter is passed there.
Bug: T156996
Bug: T157106
Change-Id: I83433cf57b6e040cdb69f3ad8807a999c4f931a5
Previously:
* Engines had to be registered in $wgRCEngines.
* The RCFeedEngine classes took no constructor arguments and
were expected to send whatever text is previously formatted
without any information about it. This generic design was
flexible in allowing one to use any formatter with any engine
with minimal configuration and no need for additional classes.
* Each feed configured their destination by setting a 'uri'
option that encodes the name of the engine in PHP as the uri
scheme. Other uri components had to be used for any other
parameters to the engine (host, port, path). While fairly
limited, it was sufficient for the default engines in core.
Changes:
* Allow feed classes to be directly associated with a feed in $wgRCFeeds
via a new 'class' option - without the indirection of 'uri' and
$wgRCEngines. All options are passed to the given class constructor.
This matches the design used elsewhere in MediaWiki. (ObjectCache,
FileRepo, FileBackend, JobQueue, LBFactory, etc.)
This means we no longer enforce a 1:1 mapping of internet protocols
to a specific feed engine, and it allows settings to be passed
without being encoded as a URI neccecarily.
Main use case for this refactor is EventBus (see I7edc4d57fa),
Interestingly, this matches the (then incorrect) documentation
written for $wgRCFeeds in 2961884b43 (which mentions an 'engine'
property that would do the same thing).
* Move the default 'omit' filters and unrestricted 'formatter' handling
to a new FormattedRCFeed class, which remains the default.
* Deprecate RecentChange::getEngine() in favour of RCFeed::factory().
* Document wgRCEngines as "@since 1.22". Follows 2961884b43, ffc71cb6af.
Change-Id: I8be497c623c5d928762e3d3406a388f4d91add9a
Use of &$this doesn't work in PHP 7.1. For callbacks to methods like
array_map() it's completely unnecessary, while for hooks we still need
to pass a reference and so we need to copy $this into a local variable.
Bug: T153505
Change-Id: I8bbb26e248cd6f213fd0e7460d6d6935a3f9e468
Since T106119, moves that overwrite a redirect produce a deletion log
entry. It should be possibleto identify such deletions within other ones.
Note that this patch only applies for actions in the future because there
is no way to easily identify these cases for the software (until now).
Bug: T145991
Change-Id: I3c006bca57351d82031c4fdce0c53f65d630b0d9
This creates a hook triggered when updating change tags, so that
extensions can take actions when tags are updated.
When tagging accompanies the action, the recent change is passed
while when tagging is subsequent to the action, the user who
performs the tagging is passed.
Bug: T118698
Change-Id: Ifb0cdc43252c4185e4f216d23b8a21fb31595a37
Most planned callers of RecentChange::addTags() only want to add a
single tag, so forcing them to create an array seems unfriendly. Plus
this lets us avoid a array_merge call.
Change-Id: If001ba9ff01a33be158e7edd5c5e4de9825fa180
Currently, extensions that add tags within the recentchanges_save
hook directly call ChangeTags::addTags. This often results in
consecutive writes to the changetags table.
This provides a addTags method to RecentChange so extensions
can simply use it when they want a tag added. And all the actual
tagging is done once at the end.
Change-Id: I8df2fd983c12632337e8d2922fa357808482338c
Simplify DB callers by just having one code path.
All but some very old code paths bothered with these.
Change-Id: Iaf7a2f83146a0ed15995f9cfc74edcf16ae5a448
This is more consistent with LoadBalancer, modern, and inclusive
of master/master mysql, NDB cluster, and MariaDB galera cluster.
The old constant is an alias now.
Change-Id: I0b37299ecb439cc446ffbe8c341365d1eef45849
If a hidden category has a member added or removed
this will only be shown to users that have the user
preference "display hidden categories".
Also added method to WikiCategoryPage to check
for hidden status.
Added unit tests for both.
Bug: T127944
Change-Id: Ic53d51391fa29d5e18fef4c42b8a275816c7ba4a
Per the docs of unserialize:
In case the passed string is not unserializeable, false is returned.
Also as we can not guarentee an array was serialized in
rc_params we cant say it will always be an array...
Quite often this is just an empty string which will
result in: "s:0:"";" (not an array)....
Change-Id: Ib22f9f7468a23ce3d4f145d47f8514d08be6c957
