is not being changed.""
This reverts commit 149e99f072.
It's not necessary to change the constructor now, the module is already
using service locator to fetch RevisionLookup and ContentHandlerFactory.
The WatchedItemStore can also be gotten from there, voiding the need for
altering the constructor now. As Daniel said in T259960#6380471 dependency
injection for API modules is good but not urgent.
Bug: T261030
Bug: T264200
Change-Id: I16aa942cc800cd66a2cd538680a02b10cb0b1bfe
This reverts commit 07e547f47c.
Reason for revert: LiquidThreads extends the ApiEditPage class,
even though it shouldn't, and thus fails when the dependencies
are not injected.
Bug: T261030
Bug: T264200
Change-Id: Ib14f8a04bb6c723aa502a47ef9ccde6fe96a0ac7
This made it impossible to correctly undo changes if revision IDs are out of
order with respect to timestamps.
Removes block of code responsible for creating this bug, and updates tests to reflect the change
Bug: T190285
Change-Id: Id5837ed958023bdbf544ae0f79e2d5e05c94cc64
This trait is not needed in ApiBase and its presence here is
proving to be problematic. See I795db12.
In this patch, the trait usage (more precisely the 'use statement')
has been removed from ApiBase and accordingly the signatures of
ApiWatchlistTrait::getWatchlistValue() and ::setWatch() have been
altered to now require User object.
With these changes, the abstract getUser() method in the trait is no
longer needed, so it has been removed also.
All core usages of the affected functions are fixed in this patch.
The trait is used in only one extension according to codesearch tool,
the extension will be fixed in Ic22e163.
Bug: T262175
Bug: T248512
Follow-up: Ia18627b9824dca81f44f0571e8420d89b7626cf6
Change-Id: Idabcea71edfca9e7ed42000a258c99ff407873d4
Ensure the content we are trying to save and the base content have
identical content models before proceeding to save so as to forestall
Exception that may be thrown by ContentHandler if it founds they're not.
There are two cases where the models are allowed to differ: Edit that
undoes content model change or edit that's meant to explicitly change
the model. The logic for these is handled separately and may succeed
or fail, but exception will not be thrown.
Bug: T255700
Change-Id: I8782732bb0fc3059693cd7035b7ebb43fd71d333
This introduces an ApiWatchlistTrait that refactors out common code
across APIs that allow you to watch pages. Some methods have been
migrated from ApiBase and changed completely, but codesearch suggests
they aren't being used outside the API modules in this patch.
Bug: T248512
Bug: T248514
Change-Id: Ia18627b9824dca81f44f0571e8420d89b7626cf6
This is to fix issues with reverts being marked incorrectly,
especially when using the undoafter param which, until this patch,
wasn't really supported.
Honestly, EditPage is such a mess that I'm not sure if this is good
or not, but I did a lot of manual testing and it seems to work fine.
WikiPage::doEditContent() now checks whether the provided original
revision really has the same content as the new revision. This was
previously the task of callers, but that doesn't make much sense to
me because:
* This task would fall on EditPage otherwise and it should be more
focused on the UI side of things, without too much worrying about
backend details.
* This would require adding another parameter to WikiPage::
doEditContent(), $undoAfterId. It would be only useful for multiple-
revision undos that are not top revisions. In all other cases it
would be the same as $originalRevId.
* An extra sanity check before applying the value to the PageUpdater
won't hurt. Who knows what crazy ideas extensions might have.
The test cases are almost identical to those written for
McrUndoAction, so I decided to group them in a single file.
This should cover most undo cases.
Bug: T256915
Change-Id: I78641b3de0a012af932ea38265b695362f1f8491
Additionally, use Article::getPage() instead of relying on
Page interface proxy method in ApiEditPage, forgotten before
Bug: T239975
Change-Id: I01696a1b11b2ea5c3d94fcb9a42beea68eb7b804
The basic idea is: The sequence `$var ?? 'default'` either uses the
value from the variable (or array element) if available, or falls back
to the default value. The resulting value is then used in the if()
condition.
if ( $var ?? true ) means the variable should default to true, if not
set.
This is mostly a style change.
In ApiEditPage the $params are impossible to not be set.
Change-Id: Id67b81744fa21fe22a2d2377259e426aab67c479
This adds a baserevid parameter for detecting edit conflicts, as an
alternative to the less precise basetimestamp parameter. This is
introduced for parity with and use by the new REST api.
Bug: T230843
Bug: T34037
Change-Id: Id7565018f66860b5c2ba688777508db1b88700ae
Notably:
* In ApiManageTags, I used a switch instead of a dynamic function name,
so that the call graph will be correct.
* In ApiImageRotate, checkTitleUserPermissions() has always returned
void, this was an error introduced in 4e6810e4a2
Change-Id: Iea22616b8e7e2e0cc804619a54f8690898b2cb82
Added:
- ContentHandlerFactory
Tests:
- PHPUnit
Changed
- Calls of changed and deprecated
- DI for some service/api
Deprecated:
- ContentHandler::* then similar to ContentHandlerFactory
- ContentHandler::getForTitle
- ContentHandler::$handlers
Bug: T235165
Change-Id: I59246938c7ad7b3e70e46c9e698708ef9bc672c6
The 'editconflict' message is used by the normal edit form, and it
expects a parameter (page name), which was never passed.
'edit-conflict' should have been used instead.
This doesn't change the error code, which is mapped in ApiMessageTrait
and it already has a case for 'edit-conflict'/'editconflict'.
Change-Id: Ic95af27a67b95310b27ec344503d0a5cd88d79a2
Some of the errors are suppressed because they're phan false positives.
The idea behind this is that they'll be fixed in a future version of
phan, and we'll just have to remove the suppressions.
Note: I'm disabling UnusedPluginSuppression so that we can start suppressing
issues even if they're still disabled. The sniff should be re-enabled
as soon as we upgrade phan.
Bug: T231636
Change-Id: I0f7fa06a9e03fbb86c7a5eb6e50a850bb258a7f7
This reverts commit 01564dcbf7.
Reason for revert: I find the empty case far clearer as to intention than this change, and I doubt this is such a hot path that optimization should take precedence over clarity. If we really want to get rid of the empty case, I'd have structured it somewhat clearly as
if (
// always allow undo via api, T230702
!( $params['undo'] > 0 ) &&
// [something else here to explain this clause]
$contentHandler->supportsDirectApiEditing() === false
) {
But even that is IMO less clear. At the very least this seems to deserve more discussion.
Bug: T230702
Change-Id: Icc16250faad31abf7308c4af9be2e7658540346c
The code in EditPage and SpecialMovepage does this primarily in web UI
code paths that aren't called by the API. EditPage also has a check in
the internal code path used by the API, but ApiEditPage runs its own
permissions check first and won't reach that code path.
Bug: T216245
Change-Id: I6263c8b60a24f3195dba583463f1ce4b004f82f5
Enforce partial blocks and display a slightly different block
notice depending on if the block is sitewide or not
Bug: T197117
Depends-On: I675316dddf272fd0d6172ecad3882160752bf780
Change-Id: I8a3635a4a04a33912eb139b7b13c4bd874183d31
Bugs fixed:
* The major one was that appendtext and prependtext seemingly didn't
work at all on empty pages in the MediaWiki namespace, because
$this->getTitle() was being used in place of $titleObj.
* Negative undo/undoafter were not rejected in a correct fashion.
* If a logged-in user who was not allowed to upload images tried to
create an image redirect, the error message used a nonexistent key.
I also replaced assertEquals with assertSame.
I also removed a bit of clearly dead code. This includes a "break"
following dieStatus(). If we actually want this break so that nobody
adds a case after the default and then removes the dieStatus() so the
switch incorrectly falls through, it could be re-added with
@codeCoverageIgnore.
I put the fixes in the same commit as the test changes because I like to
keep fixes together with the tests for those fixes.
All code now shows up as covered locally, except for one line that seems
to be a PHPUnit bug.
Change-Id: I9375bc5f40268fd681a2d447c66a03f40b23390a
It is the job of git and svn to provide this information/metadata.
The form was different, some with short, some with long month name
some with leading zero at the day, some without.
The year is also present in the Copyright clause
Change-Id: If006907b82b9e45f13cfa2e45d41107a95570e1a
PHP 7.2 broke existing functionality in making count( null ) raise a
warning. So add tests for null all over the place, or change tests where
we know the value is null or an array (but not false, empty-string, or
0) to just cast to boolean.
Bug: T182004
Change-Id: Idfe23a07daa9f60eee72f2daf04304be87057a29
Get rid of the hack that turns unicode into hexadecimal codes for
browsers that don't support unicode, and prevent their edits entirely.
And instead of relying on $wgBrowserBlacklist, use a hidden HTML form
field - if the contents are mangled and don't match the original, then
reject the edit.
Bug: T67297
Change-Id: I20c2e396d7dfd6a3b23b94b218f94a847522576b
Links generated by the API are now aware of the user's preferred
language and will show documents in that language if available.
To test, log in to mediawiki.org and set your language preference to 'es',
then on an MediaWiki installation with this patch view the generated
expanded API help at `api.php?action=help&recursivesubmodules=1&modules=main`.
Each link to documentation on mediawiki.org should take you to its
translated /es subpage, if one exists.
Bug: T104518
Change-Id: I339a1f3ae1bce9d759cf251899d57c32b1def91e
In cases where $result is non-null but doesn't have the 'spam' key,
constructing the fallback array was raising a PHP warning.
Thinking it over, we should just use a switch instead of making an array
of known fallback messages. It's longer source code, but probably less
real work than constructing an array where only one entry will actually
be used.
Bug: T153590
Change-Id: I50349c56692c4b2f9a8e560a849c36db1082120e
We already throw around some exceptions that are localized
(ErrorPageError and its subclasses, MalformedTitleException), but
there's no standard way to recognize them. Let's change that.
Then let's use them in the API to be able to have internationalized
errors when such exceptions are caught, instead of wrapping the
English-language version.
Change-Id: Iac7c90f92a889f8de9dae373547c07b884addaea
API warnings and error messages are currently hard-coded English
strings. This patch changes that.
With a few exceptions, this patch should be compatible with non-updated
extensions:
* The change to ApiBase::$messageMap will blow up anything trying to
mess with it.
* The changes to the 'ApiCheckCanExecute' hook will cause a wrong
(probably unparsed) error message to be emitted for extensions not
already using an ApiMessage. Unless they're currently broken like
Wikibase.
Bug: T37074
Bug: T47843
Depends-On: Ia2b66b57cd4eaddc30b3ffdd7b97d6ca3e02d898
Depends-On: I2e1bb975bb0045476c03ebe6cdec00259bae22ec
Depends-On: I53987bf87c48f6c00deec17a8e957d24fcc3eaa6
Depends-On: Ibf93a459eb62d30f7c70d20e91ec9faeb80d10ed
Depends-On: I3cf889811f44a15935e454dd42f081164d4a098c
Depends-On: Ieae527de86735ddcba34724730e8730fb277b99b
Depends-On: I535344c29d51521147c2a26c341dae38cec3e931
Change-Id: Iae0e2ce3bd42dd4776a9779664086119ac188412
