Don't request unused prop=info and its rvprop=watched;
use formatversion=2 instead of indexpageids to simplify processing the
returned query object.
Bug: T103981
Change-Id: I1c0f9b0245a823413f5fed78bc1d0d46ef978cef
Popups' API query request specifies the string 'true' as the value for
some boolean request parameters, but that's functionally identical to
specifying boolean true -- both get encoded as param=true in the HTTP
request.
Change-Id: Icc5f4b8ef6faaffe4a23c964f8cd8c2582d28381
This is needed for wikis that use LanguageConverter, since the title
attribute is converted to the user's variant, and the canonical title is
needed for the API unless converttitles is specified.
For filtering, instead of comparing linkHref and expectedHref, filter
out links that have other query parameters or aren't in content
namespaces.
Bug: T93605
Change-Id: I5534753307ed5e1d4b27c52c616fd143b2a397e1
Both the title and the extract were being html escaped thus producing
string like ' and " when used with .text(). So, we now use
document.createTextNode() for the normal text and .text() with the bolded
one.
Bug: T93720
Change-Id: I6bbc52e427dc636b7b0be1ad4f749d9273ff61b3
If the URL of the thumbnail has suspicious characters like ', " or \
return a <span> instead of trying to render a thumbnail.
Bug: T88171
Change-Id: Ide052ea2a7de166599d077a385a6e788bfa63302
Instead of replacing all instances of the title in the extract -
'$1<b>$2</b>$3'
We now put symbolic strings there which we use to split the string
and then make an array of text and <b> elements that get appended
to $contentbox.
Bug: T76378
Change-Id: I02222bbff84532f63cac67af1bf889c328ec6ff2
The script depends on a title being present in title attribute. So
skip links without title attributes (which were getting empty
attributes due to the implementation of removeTooltips) and bail on
trying to open a preview for links without a title.
Bug: 67728
Change-Id: I4cc744bea10af34741681f11e03d77b3d53e3a3b
removeBracketsFromText doesn't count it tracks nesting level, and
brackets are [] not (). Minor detail.
Change-Id: I8f42b1ccae5233d7b6062bb311eced6ef3085d4a
In combination with other gadgets the error
TypeError: re.query is undefined
can occur because the request contains titles:'' and the response is [].
Change-Id: I46d02e7c48a55cd9af20fd7ef729b4d26f69a8c0
Casees:
1) Hello (and welcome)
2) Hello )and welcome(
3) (Hello (and) welcome)
The current RegExp only takes care of (1). The new code should
take care of all the cases.
Bug: 65138
Change-Id: I39f38beed0cf612067d6fc61b18295d1d38fcc22
These links were not ignored due to a typo. This also threw an error,
because these links produced query.interwiki and not in query.pages.
Change-Id: I3b5c5b900209323e3e72b41e3b02e90f1f53b4eb
