Support specifying reference contents with data-mw.body.id

If data-mw.body.html is not set, search for the element
whose ID matches data-mw.body.id and use its contents.

.body.html is still used for new content: new references
use .body.html, and if a reference using .body.id is
changed, the changed HTML is output in .body.html.

Bug: T88650
Change-Id: I5d614927256d59a3ab20b0bd6e882b5cb8fcda20

modules/ve-mw/dm/nodes/ve.dm.MWReferenceNode.js

@@ -56,14 +56,24 @@ ve.dm.MWReferenceNode.static.blacklistedAnnotationTypes = [ 'link' ];
 ve.dm.MWReferenceNode.static.listKeyRegex = /^(auto|literal)\/(.*)$/;
 
 ve.dm.MWReferenceNode.static.toDataElement = function ( domElements, converter ) {
+	function getReflistItemHtml( id ) {
+		var elem = converter.getHtmlDocument().getElementById( id );
+		return elem && elem.innerHTML || '';
+	}
+
 	var dataElement,
 		mwDataJSON = domElements[0].getAttribute( 'data-mw' ),
 		mwData = mwDataJSON ? JSON.parse( mwDataJSON ) : {},
-		body = mwData.body ? mwData.body.html : '',
+		reflistItemId = mwData.body && mwData.body.id,
+		body = ( mwData.body && mwData.body.html ) ||
+			( reflistItemId && getReflistItemHtml( reflistItemId ) ) ||
+			'',
 		refGroup = mwData.attrs && mwData.attrs.group || '',
 		listGroup = this.name + '/' + refGroup,
 		autoKeyed = !mwData.attrs || mwData.attrs.name === undefined,
-		listKey = autoKeyed ? 'auto/' + converter.internalList.getNextUniqueNumber() : 'literal/' + mwData.attrs.name,
+		listKey = autoKeyed ?
+			'auto/' + converter.internalList.getNextUniqueNumber() :
+			'literal/' + mwData.attrs.name,
 		queueResult = converter.internalList.queueItemHtml( listGroup, listKey, body ),
 		listIndex = queueResult.index,
 		contentsUsed = ( body !== '' && queueResult.isNew );
@@ -81,6 +91,9 @@ ve.dm.MWReferenceNode.static.toDataElement = function ( domElements, converter )
 			contentsUsed: contentsUsed
 		}
 	};
+	if ( reflistItemId ) {
+		dataElement.attributes.refListItemId = reflistItemId;
+	}
 	return dataElement;
 };
 
@@ -143,7 +156,8 @@ ve.dm.MWReferenceNode.static.toDomElements = function ( dataElement, doc, conver
 			itemNodeWrapper
 		);
 		itemNodeHtml = itemNodeWrapper.innerHTML; // Returns '' if itemNodeWrapper is empty
-		originalHtml = ve.getProp( mwData, 'body', 'html' ) || '';
+		originalHtml = ve.getProp( mwData, 'body', 'html' ) ||
+			itemNode.getAttribute( 'originalHtml' ) || '';
 		originalHtmlWrapper.innerHTML = originalHtml;
 		// Only set body.html if itemNodeHtml and originalHtml are actually different
 		if ( !originalHtmlWrapper.isEqualNode( itemNodeWrapper ) ) {

modules/ve-mw/dm/nodes/ve.dm.MWReferencesListNode.js

@@ -60,16 +60,15 @@ ve.dm.MWReferencesListNode.static.toDataElement = function ( domElements, conver
 		}
 	};
 	if ( mwData.body && mwData.body.html ) {
+		// Process the nodes in .body.html as if they were this node's children
 		contentsDiv = domElements[0].ownerDocument.createElement( 'div' );
 		contentsDiv.innerHTML = mwData.body.html;
 		contentsData = converter.getDataFromDomClean( contentsDiv );
-		return [ referencesListData ]
+		referencesListData = [ referencesListData ]
 			.concat( contentsData )
 			.concat( [ { type: '/' + this.name } ] );
-	} else {
-		return referencesListData;
 	}
-
+	return referencesListData;
 };
 
 ve.dm.MWReferencesListNode.static.toDomElements = function ( data, doc, converter ) {

modules/ve-mw/tests/dm/ve.dm.mwExample.js

@@ -303,6 +303,8 @@ ve.dm.mwExample.MWInlineImage = {
 };
 
 ve.dm.mwExample.MWReference = {
+	// The HTML below is enriched to wrap reference contents in <span id="mw-cite-[...]">
+	// which Parsoid doesn't do yet, but T88290 asks for
 	referencesList:
 		'<ol class="references" typeof="mw:Extension/references" about="#mwt7" data-parsoid="{}"' +
 			'data-mw="{&quot;name&quot;:&quot;references&quot;,&quot;body&quot;:{' +
@@ -311,9 +313,9 @@ ve.dm.mwExample.MWReference = {
 			'&amp;quot;attrs&amp;quot;:{&amp;quot;group&amp;quot;:&amp;quot;g1&amp;quot;,&amp;quot;name&amp;quot;:&amp;quot;foo&amp;quot;}}\\&quot; ' +
 			'rel=\\&quot;dc:references\\&quot; typeof=\\&quot;mw:Extension/ref\\&quot;>' +
 			'<a href=\\&quot;#cite_note-foo-3\\&quot;>[3]</a></span>&quot;},&quot;attrs&quot;:{&quot;group&quot;:&quot;g1&quot;}}">' +
-			'<li about="#cite_note-.3A0-2" id="cite_note-.3A0-2"><span rel="mw:referencedBy"><a href="#cite_ref-.3A0_2-0">↑</a></span> Quux</li>' +
-			'<li about="#cite_note-3" id="cite_note-3"><span rel="mw:referencedBy"><a href="#cite_ref-3">↑</a></span> No name</li>' +
-			'<li about="#cite_note-foo-4" id="cite_note-foo-4"><span rel="mw:referencedBy"><a href="#cite_ref-foo_4-0">↑</a></span> Ref in refs</li>' +
+			'<li about="#cite_note-.3A0-2" id="cite_note-.3A0-2"><span rel="mw:referencedBy"><a href="#cite_ref-.3A0_2-0">↑</a></span> <span id="mw-cite-:0">Quux</span></li>' +
+			'<li about="#cite_note-3" id="cite_note-3"><span rel="mw:referencedBy"><a href="#cite_ref-3">↑</a></span> <span id="mw-cite-3">No name</span></li>' +
+			'<li about="#cite_note-foo-4" id="cite_note-foo-4"><span rel="mw:referencedBy"><a href="#cite_ref-foo_4-0">↑</a></span> <span id="mw-cite-foo">Ref in refs</span></li>' +
 			'</ol>'
 };
 
@@ -1126,7 +1128,8 @@ ve.dm.mwExample.domToDataCases = {
 					'<a href="#cite_note-bar-1">[1]</a>' +
 				'</span>' +
 				' Yay' +
-				'<span about="#mwt4" class="reference" data-mw="{&quot;name&quot;:&quot;ref&quot;,&quot;body&quot;:{&quot;html&quot;:&quot;No name&quot;},&quot;attrs&quot;:{&quot;group&quot;:&quot;g1&quot;}}" id="cite_ref-1-0" rel="dc:references" typeof="mw:Extension/ref" data-parsoid="{}">' +
+				// This reference has .body.id instead of .body.html
+				'<span about="#mwt4" class="reference" data-mw="{&quot;name&quot;:&quot;ref&quot;,&quot;body&quot;:{&quot;id&quot;:&quot;mw-cite-3&quot;},&quot;attrs&quot;:{&quot;group&quot;:&quot;g1&quot;}}" id="cite_ref-1-0" rel="dc:references" typeof="mw:Extension/ref" data-parsoid="{}">' +
 					'<a href="#cite_note-3">[g1 2]</a>' +
 				'</span>' +
 				' Quux' +
@@ -1265,17 +1268,18 @@ ve.dm.mwExample.domToDataCases = {
 					listGroup: 'mwReference/g1',
 					listKey: 'auto/0',
 					refGroup: 'g1',
-					mw: { name: 'ref', body: { html: 'No name' }, attrs: { group: 'g1' } },
-					originalMw: '{"name":"ref","body":{"html":"No name"},"attrs":{"group":"g1"}}',
+					mw: { name: 'ref', body: { id: 'mw-cite-3' }, attrs: { group: 'g1' } },
+					originalMw: '{"name":"ref","body":{"id":"mw-cite-3"},"attrs":{"group":"g1"}}',
 					childDomElements: $( '<a href="#cite_note-3">[g1 2]</a>' ).toArray(),
-					contentsUsed: true
+					contentsUsed: true,
+					refListItemId: 'mw-cite-3'
 				},
 				htmlAttributes: [
 					{
 						values: {
 							about: '#mwt4',
 							class: 'reference',
-							'data-mw': '{"name":"ref","body":{"html":"No name"},"attrs":{"group":"g1"}}',
+							'data-mw': '{"name":"ref","body":{"id":"mw-cite-3"},"attrs":{"group":"g1"}}',
 							'data-parsoid': '{}',
 							id: 'cite_ref-1-0',
 							rel: 'dc:references',