As part of the "details" syntax for subreferencing, we would want visual edits to be serialized into a wikitext-containing attribute.
The goal is to create a proof-of-concept, demonstrating how this will be implemented for the details approach.
Acceptance criteria:
- Parsoid should deliver the content of the attribute and there's a conversion from plain text in Wikitext
- VE is able to display the content in an editable surface
- VE applied edits can be saved and are converted back to the attribute in Wikitext
Out of scope
- Rendering in VE and the Parser - ideally we can connect it to existing sub-ref code though
- Directly wiring a live editing surface on the document into the attribute editor. Instead we will emulate it by editing a detached or hidden surface and then serializing that into the attribute. See the Wikitext 2017 editor for a precedent.