According to Daniel it only worked by accident, and stopped working
after de63ad823abe:
getPageByReference() used to do an opportunistic lookup by ID when given
an instance of PageIdentity -- which is correct for EventDispatcher,
but problematic in the general case, causing T296063.
The correct thing to do here is to use getPageById(), since the canonical
association between revision and page is by page ID.
Bug: T297431
Change-Id: Icc1df0c9ca5345e65ef5f8daf0815013d7db0943
This reverts commit 99b757465a.
Reason for revert: We may never need the 'behind-overlay' setting
and it is untested and probably broken.
Bug: T295816
Change-Id: I9e128862271697ece5241d0e98727174b42f54ff
* Add a N/A value for edit counts from anonymous users
* Only oversample with $wgDTSchemaEditAttemptStepOversample if the edit
is from DiscussionTools
* Consider $wgWMESchemaEditAttemptStepOversample for oversampling
Bug: T286076
Bug: T295995
Depends-On: Ieb3f6c6e1775c1ef53747c37003b17e3634d1c44
Change-Id: I91245a61dfbde8b5ec9b2893b9170cc4d73f7b0a
When our interface initialized on a page that the current user
recently edited (using the reply tool, the full-page source editor,
or any other way), check if any new automatic topic subscriptions
were added and update the interface to reflect that.
This requires doing some API requests after the page is loaded,
because adding auto-subscriptions happens asynchronously in a
DeferredUpdate (potentially after the user is already viewing
the page with their comment saved), and depends on the contents
of the edit.
(When using the reply tool, we could avoid this API request and
replicate the logic, but that's not implemented in this commit
to keep it simple.)
Bug: T284836
Change-Id: Ic0fabda0de4ebbc5e424f49641e6b03ebb4b7e6a
Usually this isn't a problem, because the comments are marked as
template-generated and we don't allow replying to them. But we had a
special case where we were trying to skip over some invisible
elements, which was causing us to skip into the middle of the
about-group in some cases. When Parsoid sees that, it serializes the
contents twice.
Bug: T290940
Change-Id: I9fe0b8d43ab874ccef371990799f77bfc46bc954
We do something similar in CommentItem.js with a moment object.
The object can be converted to a string when required.
Change-Id: Id7221e9201db0d89c3b771574634c878c9515ca0
* Use featuresEnabled.newtopictool to decide when new topic
links might exist on the page, instead of relying on
`#ca-addsection` existing. Change the logic of that feature
flag to check if __NONEWSECTIONLINK__ is on the page.
* Render the add-topic button locally in a hook to replace
the one suppressed by onMinervaNeueTalkPageOverlay. Do so
whenever the newtopictool feature is enabled (see above).
Bug: T270537
Change-Id: I3e3f7403b3b86bb84fcb75a8833919512519b70f
This should avoid them showing up in unexpected contexts where parser
output is used (e.g. API T292345, search T294168, action=render).
Also rename some variables to be directionality-neutral.
Bug: T292345
Bug: T294168
Change-Id: Ibcac44ee10f0842e205d9dd9a7f3a935ce0c690b
* Explode boundary point tuples passed to computePosition
* Note that we won't use previousSibling instead of array_reverse
* Simplify logic using xor
Change-Id: I927256e31b5e441aade91b4fd0d83d8f0d89afbe
This doesn't sort by the topic name, but the hidden
sub_id field, leading to a confusing order.
Bug: T273342
Change-Id: I6146abf05544d40c9ef0d2e8c58d020e5a5fa8a2
We always do our processing in the parser now, so we don't need the
marker comment to detect whether we've already processed the page.
Bonus: include the time taken by our processing in the limit report.
Bug: T291831
Change-Id: Ife7ddffbad1b1495b004739212002a98fdebe6c0
The PHP DOM extension measures lengths and offsets in Unicode codepoints.
Our PHP code used UTF-8 bytes, causing some offsets to be slightly off.
Now it mostly uses Unicode codepoints as well (we're forced to use bytes
in a few places, because preg_match returns offsets in bytes).
In practice, this had no visible effect to the user. It caused the
markers `<span data-mw-comment-end="..."></span>` to be placed at
the end of their container instead of the correct position when the
timestamp contained multibyte characters (e.g. "ź" in Polish); but
the correct position is usually at the end of the container anyway.
In the test cases, the only difference is placing these markers before
a trailing line break inside `<p>...</p>` tags rather than before it.
The patch also accidentally fixes another bug, where element nodes
with no children (mostly <img>) were incorrectly excluded when calling
cloneContents(), because they were treated as if they were text nodes.
Change-Id: Iccdccf1078598f4b62cab96225e9c85a4c0e93ee