This requires 1.42 for some new names
Changes to the use statements done automatically via script
Addition of missing use statements and changes to docs done manually
Change-Id: Ic1e2c9a0c891382744e4792bba1effece48e53f3
For example:
* Use the more meaningful str_contains().
* Add missing type hints.
* Make use of early returns/guard clauses.
Change-Id: Id150d1b17a80ea637a0639a8f2fd7fd017ad23b1
This was an unfortunate mistake in the refactoring in I2ccb587,
caused by incomplete documentation and a confusing mixture of
possible return types.
I9166c2b fixed one of the two places already. The situation in this
patch here cannot really happen in reality (there is nothing to
remove when the page is empty). Still I think the code is easier to
read when the two places behave the same.
Change-Id: Iea51c3a7a8185cbc3771143353f4795dde712ec4
It should fail on null but it should create the page if it doesn't
exist or doesn't have any content yet.
This is breaking the special page, see:
[[de:234828092#New_special_page_to_fight_spam_//_Neue_Spezialseite_zur_Spam-Bekämpfung]]
Change-Id: I9166c2bdcfacb4b19706d246fbf99b2f24ca4cc6
* Remove stray `@ingroup` from file blocks, move to class block.
* Fix mention of "WAN" cache where actually APCU is used.
* Document that the storage class takes a local-server cache.
This is an important requirement since the class has no
coordination for purging or other invalidation. It expects
an uncoordinated cache.
* Rename "load" to "loadConfig" as it's ambigious what it means among
the half dozen other "load*" methods in this class. Also inline
loadFromConfig and loadComputedUncached while at it to further
reduce this.
* Rename "loadConfigContent" to "fetchLatestConfig" to match
the existing fetchConfig, which does the same thing except it queries
the primary db using READ_LATEST.
* Use Html.php when building HTML, instead of legacy Xml.php.
While at it, also switch a few to Html::element instead of
Html::rawElement (aka Xml::tags) by using Message->text() for
messages that are not expected to contain rich wikitext.
Change-Id: Ic74d1597aa9201b371894e7a4bf9361752d9db21
Doing unset on array leads to the final array turning into associative array
and gets blocked by the validator.
You can check that it's broken in Persian Wikipedia, beta cluster or
localhost. Tested locally, fixes the issue.
Bug: T337431
Change-Id: Ib1be294bae1ae057dfb9a4445a8e13ac72b333b9
Regarding array building: Instead of adding to array with
$array[] = 'foo' and then doing array_flip(), simply do
$array['foo'] = true;
Regarding tests: I originally wanted to create a unit test but I ended
up mocking so many things that it wasn't worth it and the config variable
is globaly which first we need to clean up after deployment is done.
Bug: T337431
Change-Id: Iac8dca7078668ee3441d19b6aafe499c1aa0d732
This is a direct follow up for I6373fa6 where we apparently fixed
half of the cases while breaking the other half. There was actualy
a code path that can return null, and anther one that can return a
status object.
Since there is never anything done with the status object we can as
well get rid of it and always return null in case of an error.
Bug: T337431
Bug: T279275
Change-Id: I2ccb58756182897bcd6649c9f589e2f7a0321b20
We will have a pretty large list of blocked domains that we need to
swift through in each edit for any added domain. In order to cacth
subdomains being added, we have to do all sorts of complicated
operations and string search in large set of strings which is quite
slow. To fix that, let's simply pretend a user who has added
foo.bar.com, also added bar.com and com and do exact match in array of
strings making it much faster.
h/t Krinkle for the idea
Bug: T337431
Change-Id: I96795ed7d1a25f051db0b591dde21b032b138ded
It is behind a feature flag. Improvements on it can happen in follow
ups. The patch is already quite massive.
Bug: T337431
Bug: T279275
Change-Id: I3df949c4d41ce65bb4afa013da9c691ac05fc760
