Completely remove the EchoMail and EchoInteraction instruments,
supporting code, and config variables. Note well that the
EchoEventLoggingSchemas config included an entry for the
already-decommissioned Echo instrument so remove that too.
Bug: T344167
Change-Id: Ic0c44737d2c4a78ec19e67b8b8cd4e6cfb8e14fa
In EventMapperTest, suppress notifications created by calling
getExistingTestPage. So far, this test only worked because the existing
test page is created in addCoreDBData (and not inside the test), but
this will no longer be the case after core change
I308617427309815062d54c14f3438cab31b08a73. Clear the PageSaveComplete
hook handlers to prevent that.
DiscussionParser has a static cache of revision data that can become
stale when data is deleted between tests (because revision IDs can be
reassigned to different pages, similar to T344124). This cache seems
needed, and converting the class to a service seems hard, so add the
page title to the cache key to try and avoid collisions. This can still
break if two tests are using the same page, which is hopefully quite
unlikely.
Change-Id: Ic4cbd8ff424e1260544ff9754e0c89dd4bb2f733
The use of "HookHandlers" attribute in extension.json makes it possible
to inject services into hook handler classes in a future patch.
Bug: T315938
Change-Id: I84027e2899b4a013b78fe4e95f191f1e4c89b5f8
Singletons are bad, amongst other reasons, because they're never reset
in tests. They can therefore occasionally cause test failures if the
cached data stored in one of these singletons becomes stale.
As noted on the task, ideally these two classes shouldn't exist at all,
and core should be responsible for caching the information it deems
expensive to compute.
As a temporary (TM) workaround, make both classes actual services, so
that the setUp/tearDown logic in MediaWikiIntegrationTestCase can
properly reset them between tests.
Dependencies are intentionally not being injected, precisely because
these classes should just be deleted, not improved.
Bug: T344124
Change-Id: I58b8d9610f9447468235b94d25732528ab6acce6
Always set the default value to HTML, and then potentially change it in
onUserGetDefaultOptions.
Bug: T341931
Change-Id: I19efaa2361cd936125b623f90d9a564caa51d4f1
Most notably, CatchableFatalErrorException and the code that throws it
are unused since Ic5712c4ce265b6faabce7a4028b4294fe3c73f18 (in 2016).
Bug: T328220
Change-Id: I5497347e41b21d2623b7e79bf7f977268a809c1d
Pass a relevant number of unread notifications to the messages
'echo-mark-all-as-read' and 'echo-mark-wiki-as-read' each time
they are used, in order to facilitate the use of PLURAL in
the messages. This allows translators to say e.g. "mark both as
read" when there are two unread notifications, for instance.
Bug: T321462
Change-Id: Ida65be2d7f2663d9802f71c7ea1eb588db8cd254
Creates a notification type to notify users when their
user page is edited.
Co-authored-by: Kunal Mehta <legoktm@debian.org>
Co-authored-by: Ed Sanders <esanders@wikimedia.org>
Bug: T3876
Change-Id: Ibe0092a96d96f6fa9d93991418b723f3e70e1b75
Error messages included a non-existent 'code' parameter; use the
curl error message instead. Also log the error status (which is
there somewhere in 'response' but logstash doesn't manipulate JSON
fields well).
Bug: T328128
Change-Id: If295d7a3b397b22301986a226a54df908c2ccf4d
In certain circumstances, it's possible for the user account to
have been deleted by the time the deferred update runs.
Bug: T318081
Change-Id: I3d1b5d1f034bf239e25d4af484b6329582b49bdc
It would be nice for editors to be able to see the particular edit
that was their 100th, 1000th, etc. This patch changes the link of the
edit count thanks notification to the diff of the edit, rather than the
page the edit was on.
Note that this link will only work for newly created notfications. Past
edit count notification database entries do not store what revision id
the edit was on in the 'echo_events' table, so there is no way to link
to them unless the table is updated.
Bug: T326998
Change-Id: If81fd71ce6f99ad3883a3bfbfbd798270f762d37
Instead of turning the array keys into an array of strings and
searching that we can access the original array by key.
Change-Id: Id4fddd242cff66ee70fa1cc07cdcb6938482696c
Switch the callers to the standard Maintenance::waitForReplication() method.
Avoid using LoadBalancer::waitFor(), which will made internal soon.
Bug: T326274
Change-Id: Idf0562eb33d90ff70eb3a7d5d7f54da944423d27
This changes the page linked notification to link to the diff of when
the page was changed and removes the "View changes" button.
Bug: T324849
Change-Id: I25a3de3ce5c02be3abbab1d2d2dd1aad8d287cbd
The addition of the foreign key echo_push_subscription_ibfk_2
in 4abfbd3 did not affected already existing installs,
because the updater part was missing.
The addition of a foreign key also creates an index under mysql,
which is also not part of older installs and needs to be created.
Also the new field eps_topic from c188dac is missing and now added.
The new index echo_push_subscription_token from 3513c64 is added.
The drop patch set d35c502 assumes that the foreign key exists,
now every step of the update is running separately to avoid one failing
update to skip the remaing update steps.
This avoid issues on wikis installed before REL1_35 and updated with the
patches from REL1_36
Bug: T322143
Change-Id: I0759b82ad91849880c784e412e04dd53f26df6a2
The loop accidentally overwrote the whole array in every iteration.
It's not necessary to define these empty arrays at all: when working
with multi-dimensional arrays, PHP will create them as needed.
Follow-up to 4835b6b1cb.
Bug: T323491
Change-Id: I5043efc31722d2b9cd2a8dd482594355a761008e
Passing null to functions that expect a string is no longer allowed.
Here in this code it can apparently only happen in test scenarios.
That's why I use a very minimal workround that still calls the same
code. The lookup will return an empty string and trigger the expected
code path below.
Bug: T311448
Change-Id: If04924a22b82ae29b8f9327668e8965d6d2a490d
The key must match what is specified in the constructor of the
EchoNotificationDeleteJob.
The job specify their name in the constructor through __CLASS__, which
means that with the class rename this key also needs to be updated.
Bug: T322342
Change-Id: Id1eebb0cc3ddcdabe31c5531cd32f032ed52a69c
When MWEchoNotifUser::getGlobalMemcKey does not find a global user,
MWEchoNotifUser::getCountsAndTimestamps can return null for 'global' key
Just fall back to the local count/timestamp
Bug: T317330
Change-Id: Ie11b0c93c0a45fb1f873bd26788beaa3d0c6a9d7
Added PreferencesGetIconHook to add an icon on the Special:Preferences section.
Bug: T317419
Depends-On: I6d5730d47e7595b1705787995fe5db2fe734d7f5
Change-Id: Ia5eefad09f533f4654bbda123b4588f246bc1ace
This should result in no visual change to Echo, but allows
us to change the Vector implementation here.
Removes all styling when vectorvisualenhancementnext=1
is enabled.
Bug: T257143
Change-Id: Ice602ba246b51252ff0527a531aa4d5d3baee64e
EventLogging::logEvent() overwrides the $revId parameter when the schema
has been migrated to the Event Platform. Add a note and use a $revId of
-1 to make this consistent with other calls to EventLogging::logEvent()
for schemas that have been migrated to the Event Platform [0].
[0] https://codesearch.wmcloud.org/search/?q=%3A%3AlogEvent%5C(%20%27%5Cw%2B%27%2C%20-1
Bug: T318263
Change-Id: I5fa515d853326307619b044d25c2f672bcb6de63
The compare with DEFAULT_TS_DB does not work, because postgres using
different timestamp format on the database
Bug: T244898
Change-Id: I81fe88f78d94499ae9082a4fb6b1191bd432f977
I kept the "watchlist-change" and "minor-watchlist-change" rules
separate for now since while those are the same pattern the $bundleString
prefix isn't exactly the event type.
Change-Id: I360849870793d7730385dec0f57c92cdabf549d1
It is declared to be able to, but is not checked here expecting an array, resulting in `PHP Notice: Trying to access array offset on value of type bool` if false.
Change-Id: I8b3ebdd3b5e6fea9c84816b890198194951ea5c0
The code would repeatedly read primary, always getting
the same result. Let's avoid this and do the query
at most once.
Change-Id: I4c09b92cac5e4dc7fd8d36f9751be4f9149181cb
MediaWiki does not use foreign keys in the database schema
Additional changes to the schema:
- Mysql and sqlite creates index for each foreign keys, keep the index
to allow fast joins with this columns by explicit statement on new
wikis
- Explicit create statement for the UNIQUE index on column
eps_token_sha256 to make the index more visible on the sql file
- Rename existing index to match the prefix for existing indexes
Bug: T306473
Change-Id: I4bd29a6d0f9515e1a678c2a967799b90ef22f7b9
Replaced WikiPage::newFromID with newFromTitle,
because a Title object exists and could be reused
Bug: T297688
Change-Id: Ide30f259477ed8e0b48df31f5a23cafeb38d7316
Extensions using Phan need to be updated simultaneously with core due
to T308443.
Bug: T308718
Depends-On: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
Change-Id: I08d1a4122d1c8052d1cdfc6f8249abdaa80067a7
The notification icon was previously server rendered without the
`mw-ui-button` and `mw-ui-quiet` classes and then these classes were
added when the JS kicked in which resulted in a 1px layout shift as
`mw-ui-button` applied a 1px border. `mw-ui-quiet` is needed to make the
border transparent.
Change-Id: If44919798815eb6959e6dd9873db5f0bc533d742
As far as I can tell, EchoIteratorDecorator is functionally identical to
IteratorDecorator from includes/libs/iterators/ in core, so why have a
separate copy? (EchoIteratorDecorator does pass through the return value
of next() or rewind(), if any, but there shouldn't be any per the
Iterator interface.)
Change-Id: Ic763ec19c15f67d9c9b42ebffb88c52b9056ed22
Instead, prepare for UserIdentity being returned by the
UserGroupsChanged hook instead of a User
(If0b3533804aa750faed1d5f96bac98129639586d), check that the user is
local, and convert to User with a UserFactory.
Bug: T255309
Change-Id: I9e246022e7dd15ea1979dbe87b5ab0eedbbde117
Remove sql patches, mention the gerrit change to check version
- Drop field etp_user in Ib57510e6b0e9202a7e035f8ea59955dca8a0b24a
- Add index echo_notification_event in
I99fc31a62b37259d7000577583c8e8a599e400f9
- Maintenance script RemoveOrphanedEvents in
Id42a893be81176f37d9d5e3fd7e850d3a6d42e10
- Add field event_deleted in Iacd098573efd92bb1e3fcd7da4cd40cea9522f15
- Add index echo_notification_user_read_timestamp in
Iacd098573efd92bb1e3fcd7da4cd40cea9522f15
- Add index echo_target_page_page_event in
Iacd098573efd92bb1e3fcd7da4cd40cea9522f15
- Add index echo_event_page_id in
I15abd7c901df58f7eb9e5476d58a1bc12835645e
Change-Id: I227801f850bb9d60488b5dd8b1f84c4c22708f5d
The default timestamp of 00000000000000 cannot represent as timestamp,
because it gets a negative timestamp -00011130000000
This is needed for proper cross-RDBMS support
This reapply a change from I46206e0b3a687dff3168a81cf0020e669133e876,
reverted with I1c8c409b7820512b3e31246a7f3d8c1cf4db209c.
Bug: T244898
Change-Id: I109b783de0a8d60ccb161b280ce5fa09e145017b
Make buttons match the style guide a bit more closely:
* Use bold text
* Add border-radius
* Tweak padding
Change-Id: Idc929f2b7846c71c9a3011376402aaee65b32dcf
