This means the reference list is always in sync with the model's
understanding of which references are available to edit.
An exception is left in if there are no references in the model
at all, as will be the case on he.wiki, as all references are
template generated. In this case we continue to use the Parsoid
DOM, so that there is a static rendering of the reference list.
Bug: T336865
Change-Id: Iaf1089c9de532e7749c9cb70a9e697917955dca8
Minimal test case:
<ref name=a>a</ref>
<ref name=a/>
<references/>
This renders as "1. ↑ 1.0 1.1 a" in both the legacy parser as well
as Parsoid. The moment you start editing this in VisualEditor the
space before "a" disappears. This patch fixes this inconsistency.
Change-Id: Idfea1a445fc98a0433640b4f706fafcc4e236c18
html/php sections are added since otherwise it complains that the
"Test lacks html or metadata section on lines"
Change-Id: Ib1c47be09bdbe1e84b595373ad71772f2a983fc9
This fixes a minor inconsistency: A reference that comes from a
template and is already reused outside of the template is only
partially available to VE, and previewed with a warning message
because of this:
"This reference is defined in a template or other generated block,
and for now can only be previewed in source mode."
This was missing in the reuse dialog.
Note this patch is not meant to make any design decision, but to use
the existing design consistently.
You can test this with and without the Citoid extension. It works in
both cases.
Bug: T336372
Change-Id: I962cf111b1882bcd736f1090ca17d2b176495d2f
This fixes a minor regression introduced in Ib003b8a. The problem is
that undefined is not equal to anything in JavaScript, and not smaller
than 2 either.
Bug: T241885
Bug: T335410
Change-Id: Ia6deb291d923b88a08ceac8fbc0efb682e14f358
This patch also updates a second place that does almost the same.
That other place also excludes placeholders. We intentionally don't
do this in this new place.
Bug: T241885
Change-Id: Ib003b8a7bbe247db6f7da0a4efcfd4e5967fd033
This issue was introduced in I7a8ec9f. I tried to understand what this
does but can't see an effect. The behavior and appearence seems to be
identical with and without this code. In case it turns out this is dead
code (e.g. not needed any more with our current browser requirements)
it should be removed. Otherwise the selector fixed.
Change-Id: I0bfae951aa4543c528dd7e04c00a0e978f1ce49e
* leadingWS is always '' in all the uses so far and it originally
came from Cite when we first extracted the extension API.
* It is unnecessarily confusing and serves no real purpose.
Change-Id: I5f2f0f123dc6938735c09b96a42c1923b0a49085
