introduce seperate content property on activeTag

Author: elpoelma

packages/ember-rdfa-editor/src/utils/_private/rdfa-parser/active-tag.ts

@@ -25,6 +25,7 @@ export interface IActiveTag<N> {
   predicates?: ModelNamedNode<N>[] | null;
   object?: ModelNamedNode<N> | ModelBlankNode<N> | boolean | null;
   text?: string[] | null;
+  content?: string | null; // contains `text` concatenated, or the `content` attribute
   vocab?: string;
   language?: string;
   datatype?: ModelNamedNode<N>;

packages/ember-rdfa-editor/src/utils/_private/rdfa-parser/rdfa-parser.ts

@@ -979,11 +979,7 @@ export class RdfaParser<N> {
     attributes: Record<string, string>,
     predicateAttribute = 'property',
   ) => {
-    const textSegments: string[] = activeTag.text || [];
-    const object = this.util.createLiteral(
-      attributes['content'] ?? textSegments.join(''),
-      activeTag,
-    );
+    const object = this.util.createLiteral(activeTag.content ?? '', activeTag);
     this.contentNodeMapping.set(node, {
       subject: sayDataFactory.resourceNode(
         this.util.getResourceOrBaseIri(unwrap(activeTag.subject), activeTag)
@@ -1049,6 +1045,14 @@ export class RdfaParser<N> {
     const parentTag: IActiveTag<N> =
       this.activeTagStack[this.activeTagStack.length - 2];
 
+    let textSegments: string[] = activeTag.text || [];
+    if (activeTag.collectChildTags && parentTag.collectChildTags) {
+      // If we are inside an XMLLiteral child that also has RDFa content, ignore the tag name that was collected.
+      textSegments = textSegments.slice(1);
+    }
+    activeTag.content =
+      activeTag.attributes['content'] ?? textSegments.join('');
+
     if (
       !(
         activeTag.collectChildTags &&
@@ -1120,8 +1124,7 @@ export class RdfaParser<N> {
 
         // Reset text, unless the parent is also collecting text
         if (!parentTag.predicates) {
-          // Can we simply remove this statement?
-          // activeTag.text = null;
+          activeTag.text = null;
         }
       }