This ensures that each repository's "npm test" command is fully
independent of wherever it might be in the filesystem.
Bug: T206485
Change-Id: I54d052bed4ceed175a938896154fd6ad5c89f964
We can resolve this bug by either replacing the bogus "return false"
with the intended "return [ false, … ]". Or rely on the code a few
lines below that also bails out with a "return [ false, … ]" when to
many parameters ($cnt is not 0 then) are present. The tests prove both
solutions are equally valid.
Bug: T211576
Change-Id: Iadd55c134dede7042cfd152c69bc8f27b59d8912
The biggest issue with this code was that it was tracking the exact
same state in two ways: Processed array elements got removed from the
$argv array, *and* the $cnt was decremented the same time. This is not
necessary and a potential source of confusion and errors.
I carefully transformed the code. I'm sure it still does exactly what
it did before. The tests should prove this.
Change-Id: I642d38e7944aa3e2239179fa58e1e231b4698263
This gives the code a little more structure without changing anything
it does. A section is extracted as a named function, and some lines
are moved and bundled where it made sense.
Change-Id: I51909517021bee9dc618efe5fbe40adfc29dc6af
This is split from I642d38e and does nothing but adding test cases
that document the current (broken) behavior.
Bug: T211576
Change-Id: Iee313d26e7bed6deb34101e37736a1c697947905
The code before was not wrong, just deeply nested. The worst case was
that the final $upArrowLink.attr( … ) might have been called on zero
elements. jQuery is fine with this.
Change-Id: I62e7286c7fe906544fe148e1122c60cc8db070f3
This code was written in 2014 and it seems that some change to
ve.ui.MWTemplateDialog since then made it not work. No idea why or
when it broke. But we can implement the functionality much more easily
now by overriding #setApplicableStatus, which works correctly.
Bug: T210796
Change-Id: I6e87ca979a96785fc74f1a57f7d80ba4b43705f2
This sets the label to be the same as the default value inherited from
ve.ui.MWTemplateDialog. Looks like it's no longer needed since change
Ia8fb88d3501ffa2c26add4419da5463a926f45d1 in VE-MW (2014).
Change-Id: Icd64d22aef29a3791352e1839993ea8462a147f3
The method #hasUsefulParameter has been accidentally made to always
return false in fdde5b71af.
Bug: T210686
Change-Id: I215924c2f693842abd8026404d17c725192272c3
The class="mw-cite-backlink" is part of a message. It should be
considered code and not be customized, but can be. Frwiki for example
localized it.
The new code still hopes to find the class, and still hopes the first
child is the text node containing the plain text up arrow. (See the
message "cite_references_link_many" which contains all this.) This code
path is kept because it is more performant.
If the class is not found, the upwards traversal done via .closest()
stops at the <li>. It then traverses down into the firstChild nodes,
hoping to find the plain text up arrow this way.
This patch makes the code work with the customizations found in frwiki.
Bug: T205270
Bug: T210508
Change-Id: I32552ebe820ee12aea1a75aa17af11298dc7536a
This change makes the code much, much more robust. See, almost all HTML
relevant for this feature is encoded in messages. This allows unexpected
customizations that add additional HTML elements, change the order of DOM
elements, even remove class names and elements.
The <ol class="references"> is the only HTML snippet generated via code,
guaranteed to be there, and used as an entry point because of this.
Instead of the selector utilizing a "*" to detect references with "one"
vs. "many" backlinks, we check if a second backlink exists.
The .first() copies the browsers behavior to only respect the first anchor
in case an id="…" appears multiple times on a page.
The additional .length check further down is a missing sanity check,
currently relevant on frwiki where the expected class="mw-cite-backlink"
got localized.
Bug: T205270
Bug: T210520
Change-Id: Iba9aebfd01508b283933964cfb986d7239d4cf38
