Cite's MWDocumentReferences included a Document, which caused
auto-save serialization to fail silently on the cyclical reference.
This patch moves the caching to per-instance scope, which was the
better choice to begin with. The document refs are recalculated after
each change anyway, and caching is only necessary to work around an
awkward control flow in which each ReferenceNode is responsible for
its footnote number.
Bug: T377484
Change-Id: Iebe4befd603fe07d9da2a5bb6126228ec4d5dfea
Reverses the previous logic which traversed up from a fragment to get
the full document's refs. Much other code in VE isn't ready for this
behavior, for example we can see list-defined refs but not inline refs
defined outside of the fragment.
This patch will ensure that we're only looking at refs accessible from
the current fragment, and prevents caching on fragments because the
cache uses `persistentStorage`, which is shared between fragments and
their parent document.
Bug: T374068
Change-Id: Ia38098f8b3e5a9d24c2206e11edab37d60209225
Pushes per-group knowledge down into a structured object and give it
an interface, separated from the singleton cache across all groups.
Also changes the behavior of orphaned subrefs so that they're still
rendered as subrefs, with an error placeholder where the parent
should be.
Bug: T372871
Change-Id: I84e679a8365f3fbfabaf344d99f56f6d069c0776
Singleton always steps up to the original document from fragments, to
give absolute numbering.
Bug: T370874
Change-Id: I0353649289f6c8fe26fa6bdff5d2367b7b575bac
Some wikis prefer to use Western Arabic numerals rather than the
digit transformation table associated with the content language
script. For these, we should respect the existing
$wgTranslateNumerals configuration settings.
Bug: T370585
Change-Id: I663ae28c5b5d4fdd66b8590eb0b453ec25c2db84
Significant change is that footnote marker numbers are changed from
being a CSS-rendered marker to using the rendered "fallback" text.
This could be avoided using the same trick as is implemented for the
reflist: send an inline style variable with the marker content; but
let's only do this if really necessary for user experience.
Template-produced refs are still wrong, but this patch continues to
render them as they come from Parsoid, in the content script.
TODO in later patches:
* reuse of a subref is rendered as '3.2.1' in reader view but '3.0'
and '3.1' in the editor.
* subref numbering is backwards in RTL languages
Bug: T247921
Change-Id: Ieff73769f8ebbc3724f6a9b498487c4e7d09aa2e
Citoid uses placeholder refs to preview where a new/re-used ref
would be added. This also influences the counting and preview in
the reference list for the moment the placeholder is visible.
I can image this that this might become a feature in the future
but for the moment it's a distraction IMO.
Bug: T247921
Change-Id: I5c5e84ae4b183f99530fda0736a58139e9e25d1a
This isn't the ideal solution since it doesn't exactly match the
rendered reader view, but it's a reasonable workaround and an
improvement on "undefined" numbering.
Bug: T247921
Change-Id: Ic0d88123d50e2fcb7f25e897280dbfdb6d494501
Pure refactor which shouldn't change output in production. Switches
to interfacing with MWDocumentReferences to get refs in index order.
Temporarily suppresses any subrefs, we only show top-level refs.
Bug: T247921
Change-Id: I9c8347b064173027f436722c87e15e0381c958bd
