@usejunior/docx-core 0.9.1 → 0.11.0

package/dist/primitives/schema-corpus-capture.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"schema-corpus-capture.js","sourceRoot":"","sources":["../../src/primitives/schema-corpus-capture.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,SAAS,EAAE,aAAa,EAAE,MAAM,SAAS,CAAC;AACnD,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAEjC;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,UAAU,8BAA8B,CAAC,GAAW;IACxD,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,qBAAqB,CAAC;IAC9C,IAAI,CAAC,GAAG;QAAE,OAAO;IACjB,MAAM,MAAM,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAC3E,SAAS,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACpC,aAAa,CAAC,IAAI,CAAC,GAAG,EAAE,YAAY,MAAM,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC;AAC1D,CAAC"}

package/dist/primitives/sectPrAudit.d.ts

@@ -0,0 +1,19 @@
+export type SectPrIssueType = 'missing_body' | 'multiple_body_level_sectpr' | 'body_level_sectpr_not_last' | 'sectpr_invalid_parent' | 'sectpr_in_ppr_without_paragraph_parent' | 'sectpr_reference_missing_rid' | 'sectpr_reference_dangling_rid';
+export interface SectPrAuditIssue {
+    type: SectPrIssueType;
+    path: string;
+    message: string;
+    rid?: string;
+}
+export interface SectPrAuditSummary {
+    ok: boolean;
+    issues: SectPrAuditIssue[];
+    stats: {
+        bodyLevelSectPrCount: number;
+        paragraphLevelSectPrCount: number;
+        totalSectPrCount: number;
+        referenceCount: number;
+    };
+}
+export declare function auditSectPr(documentXml: string, documentRelsXml?: string | null): SectPrAuditSummary;
+//# sourceMappingURL=sectPrAudit.d.ts.map

package/dist/primitives/sectPrAudit.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"sectPrAudit.d.ts","sourceRoot":"","sources":["../../src/primitives/sectPrAudit.ts"],"names":[],"mappings":"AAIA,MAAM,MAAM,eAAe,GACvB,cAAc,GACd,4BAA4B,GAC5B,4BAA4B,GAC5B,uBAAuB,GACvB,wCAAwC,GACxC,8BAA8B,GAC9B,+BAA+B,CAAC;AAEpC,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,eAAe,CAAC;IACtB,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,EAAE,MAAM,CAAC;IAChB,GAAG,CAAC,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,kBAAkB;IACjC,EAAE,EAAE,OAAO,CAAC;IACZ,MAAM,EAAE,gBAAgB,EAAE,CAAC;IAC3B,KAAK,EAAE;QACL,oBAAoB,EAAE,MAAM,CAAC;QAC7B,yBAAyB,EAAE,MAAM,CAAC;QAClC,gBAAgB,EAAE,MAAM,CAAC;QACzB,cAAc,EAAE,MAAM,CAAC;KACxB,CAAC;CACH;AA8DD,wBAAgB,WAAW,CAAC,WAAW,EAAE,MAAM,EAAE,eAAe,CAAC,EAAE,MAAM,GAAG,IAAI,GAAG,kBAAkB,CAoHpG"}

package/dist/primitives/sectPrAudit.js

@@ -0,0 +1,165 @@
+import { parseXml } from './xml.js';
+import { childElements } from './dom-helpers.js';
+import { OOXML } from './namespaces.js';
+function elementSiblingIndex(node) {
+    const parent = node.parentNode;
+    if (!parent)
+        return 1;
+    let idx = 0;
+    for (const sibling of childElements(parent)) {
+        if (sibling.tagName === node.tagName) {
+            idx++;
+        }
+        if (sibling === node) {
+            return idx;
+        }
+    }
+    return idx || 1;
+}
+function nodePath(node) {
+    const parts = [];
+    let current = node;
+    while (current) {
+        parts.push(`${current.tagName}[${elementSiblingIndex(current)}]`);
+        const parentNode = current.parentNode;
+        if (!parentNode || parentNode.nodeType !== 1) {
+            break;
+        }
+        current = parentNode;
+    }
+    return parts.reverse().join('/');
+}
+function collectRelationshipIds(documentRelsXml) {
+    if (!documentRelsXml) {
+        return new Set();
+    }
+    try {
+        const relDoc = parseXml(documentRelsXml);
+        const ids = new Set();
+        const relationships = relDoc.getElementsByTagName('Relationship');
+        for (let i = 0; i < relationships.length; i++) {
+            const rel = relationships.item(i);
+            const id = rel?.getAttribute('Id');
+            if (id) {
+                ids.add(id);
+            }
+        }
+        return ids;
+    }
+    catch {
+        return new Set();
+    }
+}
+function getRid(ref) {
+    return (ref.getAttribute('r:id') ??
+        ref.getAttributeNS(OOXML.R_NS, 'id') ??
+        ref.getAttribute('id') ??
+        undefined);
+}
+export function auditSectPr(documentXml, documentRelsXml) {
+    const issues = [];
+    const relIds = collectRelationshipIds(documentRelsXml);
+    const doc = parseXml(documentXml);
+    const body = doc.getElementsByTagName('w:body').item(0);
+    if (!body) {
+        return {
+            ok: false,
+            issues: [
+                {
+                    type: 'missing_body',
+                    path: 'w:document',
+                    message: 'Missing w:body element',
+                },
+            ],
+            stats: {
+                bodyLevelSectPrCount: 0,
+                paragraphLevelSectPrCount: 0,
+                totalSectPrCount: 0,
+                referenceCount: 0,
+            },
+        };
+    }
+    const bodyChildren = childElements(body);
+    const bodyLevelSectPrNodes = bodyChildren.filter((child) => child.tagName === 'w:sectPr');
+    if (bodyLevelSectPrNodes.length > 1) {
+        for (const sectPr of bodyLevelSectPrNodes) {
+            issues.push({
+                type: 'multiple_body_level_sectpr',
+                path: nodePath(sectPr),
+                message: 'Multiple body-level w:sectPr elements found; expected at most one final w:sectPr',
+            });
+        }
+    }
+    if (bodyLevelSectPrNodes.length > 0) {
+        const lastChild = bodyChildren[bodyChildren.length - 1];
+        for (const sectPr of bodyLevelSectPrNodes) {
+            if (sectPr !== lastChild) {
+                issues.push({
+                    type: 'body_level_sectpr_not_last',
+                    path: nodePath(sectPr),
+                    message: 'Body-level w:sectPr is not the final direct child of w:body',
+                });
+            }
+        }
+    }
+    const sectPrNodes = Array.from(doc.getElementsByTagName('w:sectPr'));
+    let paragraphLevelSectPrCount = 0;
+    let referenceCount = 0;
+    for (const sectPr of sectPrNodes) {
+        const parent = sectPr.parentNode;
+        const parentTag = parent && parent.nodeType === 1 ? parent.tagName : '';
+        if (parentTag === 'w:pPr') {
+            paragraphLevelSectPrCount++;
+            const grand = parent.parentNode;
+            const grandTag = grand && grand.nodeType === 1 ? grand.tagName : '';
+            if (grandTag !== 'w:p') {
+                issues.push({
+                    type: 'sectpr_in_ppr_without_paragraph_parent',
+                    path: nodePath(sectPr),
+                    message: 'w:sectPr in w:pPr does not have w:p as parent',
+                });
+            }
+        }
+        else if (parentTag !== 'w:body') {
+            issues.push({
+                type: 'sectpr_invalid_parent',
+                path: nodePath(sectPr),
+                message: `w:sectPr has invalid parent '${parentTag || '(none)'}'`,
+            });
+        }
+        for (const child of childElements(sectPr)) {
+            if (child.tagName !== 'w:headerReference' && child.tagName !== 'w:footerReference') {
+                continue;
+            }
+            referenceCount++;
+            const rid = getRid(child);
+            if (!rid) {
+                issues.push({
+                    type: 'sectpr_reference_missing_rid',
+                    path: nodePath(child),
+                    message: `${child.tagName} is missing r:id`,
+                });
+                continue;
+            }
+            if (relIds.size > 0 && !relIds.has(rid)) {
+                issues.push({
+                    type: 'sectpr_reference_dangling_rid',
+                    path: nodePath(child),
+                    message: `${child.tagName} references missing relationship id '${rid}'`,
+                    rid,
+                });
+            }
+        }
+    }
+    return {
+        ok: issues.length === 0,
+        issues,
+        stats: {
+            bodyLevelSectPrCount: bodyLevelSectPrNodes.length,
+            paragraphLevelSectPrCount,
+            totalSectPrCount: sectPrNodes.length,
+            referenceCount,
+        },
+    };
+}
+//# sourceMappingURL=sectPrAudit.js.map

package/dist/primitives/sectPrAudit.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"sectPrAudit.js","sourceRoot":"","sources":["../../src/primitives/sectPrAudit.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAC;AACpC,OAAO,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAC;AACjD,OAAO,EAAE,KAAK,EAAE,MAAM,iBAAiB,CAAC;AA6BxC,SAAS,mBAAmB,CAAC,IAAa;IACxC,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC;IAC/B,IAAI,CAAC,MAAM;QAAE,OAAO,CAAC,CAAC;IACtB,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,KAAK,MAAM,OAAO,IAAI,aAAa,CAAC,MAAiB,CAAC,EAAE,CAAC;QACvD,IAAI,OAAO,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;YACrC,GAAG,EAAE,CAAC;QACR,CAAC;QACD,IAAI,OAAO,KAAK,IAAI,EAAE,CAAC;YACrB,OAAO,GAAG,CAAC;QACb,CAAC;IACH,CAAC;IACD,OAAO,GAAG,IAAI,CAAC,CAAC;AAClB,CAAC;AAED,SAAS,QAAQ,CAAC,IAAa;IAC7B,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAmB,IAAI,CAAC;IACnC,OAAO,OAAO,EAAE,CAAC;QACf,KAAK,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,OAAO,IAAI,mBAAmB,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QAClE,MAAM,UAAU,GAAgB,OAAO,CAAC,UAAU,CAAC;QACnD,IAAI,CAAC,UAAU,IAAI,UAAU,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YAC7C,MAAM;QACR,CAAC;QACD,OAAO,GAAG,UAAqB,CAAC;IAClC,CAAC;IACD,OAAO,KAAK,CAAC,OAAO,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACnC,CAAC;AAED,SAAS,sBAAsB,CAAC,eAA0C;IACxE,IAAI,CAAC,eAAe,EAAE,CAAC;QACrB,OAAO,IAAI,GAAG,EAAU,CAAC;IAC3B,CAAC;IAED,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,QAAQ,CAAC,eAAe,CAAC,CAAC;QACzC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;QAC9B,MAAM,aAAa,GAAG,MAAM,CAAC,oBAAoB,CAAC,cAAc,CAAC,CAAC;QAClE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC9C,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAClC,MAAM,EAAE,GAAG,GAAG,EAAE,YAAY,CAAC,IAAI,CAAC,CAAC;YACnC,IAAI,EAAE,EAAE,CAAC;gBACP,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACd,CAAC;QACH,CAAC;QACD,OAAO,GAAG,CAAC;IACb,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,GAAG,EAAU,CAAC;IAC3B,CAAC;AACH,CAAC;AAED,SAAS,MAAM,CAAC,GAAY;IAC1B,OAAO,CACL,GAAG,CAAC,YAAY,CAAC,MAAM,CAAC;QACxB,GAAG,CAAC,cAAc,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC;QACpC,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC;QACtB,SAAS,CACV,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,WAAW,CAAC,WAAmB,EAAE,eAA+B;IAC9E,MAAM,MAAM,GAAuB,EAAE,CAAC;IACtC,MAAM,MAAM,GAAG,sBAAsB,CAAC,eAAe,CAAC,CAAC;IAEvD,MAAM,GAAG,GAAG,QAAQ,CAAC,WAAW,CAAC,CAAC;IAClC,MAAM,IAAI,GAAG,GAAG,CAAC,oBAAoB,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,CAAmB,CAAC;IAE1E,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO;YACL,EAAE,EAAE,KAAK;YACT,MAAM,EAAE;gBACN;oBACE,IAAI,EAAE,cAAc;oBACpB,IAAI,EAAE,YAAY;oBAClB,OAAO,EAAE,wBAAwB;iBAClC;aACF;YACD,KAAK,EAAE;gBACL,oBAAoB,EAAE,CAAC;gBACvB,yBAAyB,EAAE,CAAC;gBAC5B,gBAAgB,EAAE,CAAC;gBACnB,cAAc,EAAE,CAAC;aAClB;SACF,CAAC;IACJ,CAAC;IAED,MAAM,YAAY,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACzC,MAAM,oBAAoB,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,KAAK,UAAU,CAAC,CAAC;IAE1F,IAAI,oBAAoB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpC,KAAK,MAAM,MAAM,IAAI,oBAAoB,EAAE,CAAC;YAC1C,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,4BAA4B;gBAClC,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC;gBACtB,OAAO,EAAE,kFAAkF;aAC5F,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,IAAI,oBAAoB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpC,MAAM,SAAS,GAAG,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACxD,KAAK,MAAM,MAAM,IAAI,oBAAoB,EAAE,CAAC;YAC1C,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;gBACzB,MAAM,CAAC,IAAI,CAAC;oBACV,IAAI,EAAE,4BAA4B;oBAClC,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC;oBACtB,OAAO,EAAE,6DAA6D;iBACvE,CAAC,CAAC;YACL,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,oBAAoB,CAAC,UAAU,CAAC,CAAC,CAAC;IACrE,IAAI,yBAAyB,GAAG,CAAC,CAAC;IAClC,IAAI,cAAc,GAAG,CAAC,CAAC;IAEvB,KAAK,MAAM,MAAM,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,MAAM,GAAG,MAAM,CAAC,UAAU,CAAC;QACjC,MAAM,SAAS,GAAG,MAAM,IAAI,MAAM,CAAC,QAAQ,KAAK,CAAC,CAAC,CAAC,CAAE,MAAkB,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;QAErF,IAAI,SAAS,KAAK,OAAO,EAAE,CAAC;YAC1B,yBAAyB,EAAE,CAAC;YAC5B,MAAM,KAAK,GAAI,MAAkB,CAAC,UAAU,CAAC;YAC7C,MAAM,QAAQ,GAAG,KAAK,IAAI,KAAK,CAAC,QAAQ,KAAK,CAAC,CAAC,CAAC,CAAE,KAAiB,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;YACjF,IAAI,QAAQ,KAAK,KAAK,EAAE,CAAC;gBACvB,MAAM,CAAC,IAAI,CAAC;oBACV,IAAI,EAAE,wCAAwC;oBAC9C,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC;oBACtB,OAAO,EAAE,+CAA+C;iBACzD,CAAC,CAAC;YACL,CAAC;QACH,CAAC;aAAM,IAAI,SAAS,KAAK,QAAQ,EAAE,CAAC;YAClC,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,uBAAuB;gBAC7B,IAAI,EAAE,QAAQ,CAAC,MAAM,CAAC;gBACtB,OAAO,EAAE,gCAAgC,SAAS,IAAI,QAAQ,GAAG;aAClE,CAAC,CAAC;QACL,CAAC;QAED,KAAK,MAAM,KAAK,IAAI,aAAa,CAAC,MAAM,CAAC,EAAE,CAAC;YAC1C,IAAI,KAAK,CAAC,OAAO,KAAK,mBAAmB,IAAI,KAAK,CAAC,OAAO,KAAK,mBAAmB,EAAE,CAAC;gBACnF,SAAS;YACX,CAAC;YAED,cAAc,EAAE,CAAC;YACjB,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;YAC1B,IAAI,CAAC,GAAG,EAAE,CAAC;gBACT,MAAM,CAAC,IAAI,CAAC;oBACV,IAAI,EAAE,8BAA8B;oBACpC,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC;oBACrB,OAAO,EAAE,GAAG,KAAK,CAAC,OAAO,kBAAkB;iBAC5C,CAAC,CAAC;gBACH,SAAS;YACX,CAAC;YAED,IAAI,MAAM,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;gBACxC,MAAM,CAAC,IAAI,CAAC;oBACV,IAAI,EAAE,+BAA+B;oBACrC,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC;oBACrB,OAAO,EAAE,GAAG,KAAK,CAAC,OAAO,wCAAwC,GAAG,GAAG;oBACvE,GAAG;iBACJ,CAAC,CAAC;YACL,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO;QACL,EAAE,EAAE,MAAM,CAAC,MAAM,KAAK,CAAC;QACvB,MAAM;QACN,KAAK,EAAE;YACL,oBAAoB,EAAE,oBAAoB,CAAC,MAAM;YACjD,yBAAyB;YACzB,gBAAgB,EAAE,WAAW,CAAC,MAAM;YACpC,cAAc;SACf;KACF,CAAC;AACJ,CAAC"}

package/dist/primitives/semantic_tags.d.ts

@@ -12,6 +12,13 @@ export declare function stripFontTags(text: string): string;
  * `<highlight>`, `<highlighting>`, `<a href="...">`, `</a>`, `<font ...>`,
  * `</font>`, `<header>`, `</header>`, `<RunInHeader>`, `</RunInHeader>`,
  * `<definition>`, `</definition>`.
+ *
+ * The replacement runs to a fixpoint (loop until the string stops changing) rather than a
+ * single pass: removing one tag can splice two halves together into a *new* tag occurrence
+ * (e.g. `<b<b>i>` → `<bi>`), which a single pass would leave behind. Looping closes that
+ * `js/incomplete-multi-character-sanitization` gap. Well-formed document `tagged_text`
+ * (non-nested, known tags only) reaches the fixpoint on the first pass, so this is a no-op
+ * for real input.
  */
 export declare function stripAllInlineTags(text: string): string;
 //# sourceMappingURL=semantic_tags.d.ts.map

package/dist/primitives/semantic_tags.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"semantic_tags.d.ts","sourceRoot":"","sources":["../../src/primitives/semantic_tags.ts"],"names":[],"mappings":"AAIA,eAAO,MAAM,aAAa,cAAc,CAAC;AAQzC,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEvD;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAIxD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEtD;AAED,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAEvD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAOtD;AAED,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAMvD;AAOD,wBAAgB,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEjD;AAED,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAGlD;AAOD;;;;;GAKG;AACH,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAGvD"}
+{"version":3,"file":"semantic_tags.d.ts","sourceRoot":"","sources":["../../src/primitives/semantic_tags.ts"],"names":[],"mappings":"AAIA,eAAO,MAAM,aAAa,cAAc,CAAC;AAQzC,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEvD;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAIxD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEtD;AAED,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAEvD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAQtD;AAED,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAMvD;AAOD,wBAAgB,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAEjD;AAED,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAGlD;AAYD;;;;;;;;;;;;GAYG;AACH,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CASvD"}

package/dist/primitives/semantic_tags.js

@@ -22,13 +22,14 @@ export function stripHyperlinkTags(text) {
 }
 export function hasHighlightTags(text) {
     return (text.includes(`<${HIGHLIGHT_TAG}>`) ||
+        text.includes(`<${HIGHLIGHT_TAG} `) || // full-mode attributed form: <highlight color="...">
         text.includes(`</${HIGHLIGHT_TAG}>`) ||
         text.includes('<highlighting>') ||
         text.includes('</highlighting>'));
 }
 export function stripHighlightTags(text) {
     return text
-        .replaceAll(new RegExp(`<${HIGHLIGHT_TAG}>`, 'g'), '')
+        .replaceAll(new RegExp(`<${HIGHLIGHT_TAG}\\b[^>]*>`, 'g'), '')
         .replaceAll(new RegExp(`</${HIGHLIGHT_TAG}>`, 'g'), '')
         .replaceAll(/<highlighting>/g, '')
         .replaceAll(/<\/highlighting>/g, '');
@@ -44,15 +45,33 @@ export function stripFontTags(text) {
     return text.replace(FONT_OPEN_RE, '').replace(FONT_CLOSE_RE, '');
 }
 // General-purpose inline tag stripper ──────────────────────────────
-const ALL_INLINE_TAGS_RE = /<\/?(?:b|i|u|highlight|highlighting|a|font|header|RunInHeader|definition)(?:\s[^>]*)?>|<a\s+href="[^"]*">/g;
+// A word boundary (`\b`) after the tag name keeps `<beta>` from matching `<b>` while letting
+// `[^>]*>` consume any attributes linearly. Two earlier constructs were dropped: the
+// `(?:\s[^>]*)?` group (its `\s` overlaps `[^>]`, the ambiguity CodeQL flags as polynomial
+// ReDoS — `js/polynomial-redos` — on uncontrolled document text) and the trailing
+// `|<a\s+href="[^"]*">` alternative (already covered by the `a` name + `[^>]*`).
+const ALL_INLINE_TAGS_RE = /<\/?(?:b|i|u|highlight|highlighting|a|font|header|RunInHeader|definition)\b[^>]*>/g;
 /**
  * Strip ALL known inline tags from text. Handles `<b>`, `<i>`, `<u>`,
  * `<highlight>`, `<highlighting>`, `<a href="...">`, `</a>`, `<font ...>`,
  * `</font>`, `<header>`, `</header>`, `<RunInHeader>`, `</RunInHeader>`,
  * `<definition>`, `</definition>`.
+ *
+ * The replacement runs to a fixpoint (loop until the string stops changing) rather than a
+ * single pass: removing one tag can splice two halves together into a *new* tag occurrence
+ * (e.g. `<b<b>i>` → `<bi>`), which a single pass would leave behind. Looping closes that
+ * `js/incomplete-multi-character-sanitization` gap. Well-formed document `tagged_text`
+ * (non-nested, known tags only) reaches the fixpoint on the first pass, so this is a no-op
+ * for real input.
  */
 export function stripAllInlineTags(text) {
-    ALL_INLINE_TAGS_RE.lastIndex = 0;
-    return text.replace(ALL_INLINE_TAGS_RE, '');
+    let current = text;
+    let previous;
+    do {
+        previous = current;
+        ALL_INLINE_TAGS_RE.lastIndex = 0;
+        current = current.replace(ALL_INLINE_TAGS_RE, '');
+    } while (current !== previous);
+    return current;
 }
 //# sourceMappingURL=semantic_tags.js.map

package/dist/primitives/semantic_tags.js.map

@@ -1 +1 @@
-{"version":3,"file":"semantic_tags.js","sourceRoot":"","sources":["../../src/primitives/semantic_tags.ts"],"names":[],"mappings":"AAAA,sDAAsD;AACtD,EAAE;AACF,mFAAmF;AAEnF,MAAM,CAAC,MAAM,aAAa,GAAG,WAAW,CAAC;AAEzC,qEAAqE;AAErE,MAAM,iBAAiB,GAAG,aAAa,CAAC;AACxC,MAAM,iBAAiB,GAAG,qBAAqB,CAAC;AAChD,MAAM,kBAAkB,GAAG,QAAQ,CAAC;AAEpC,MAAM,UAAU,iBAAiB,CAAC,IAAY;IAC5C,OAAO,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AACtC,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,IAAY;IAC9C,kDAAkD;IAClD,iBAAiB,CAAC,SAAS,GAAG,CAAC,CAAC;IAChC,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC;AAC7C,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,OAAO,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;AACvD,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC;AAC7E,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,OAAO,CACL,IAAI,CAAC,QAAQ,CAAC,IAAI,aAAa,GAAG,CAAC;QACnC,IAAI,CAAC,QAAQ,CAAC,KAAK,aAAa,GAAG,CAAC;QACpC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC;QAC/B,IAAI,CAAC,QAAQ,CAAC,iBAAiB,CAAC,CACjC,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,IAAI;SACR,UAAU,CAAC,IAAI,MAAM,CAAC,IAAI,aAAa,GAAG,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC;SACrD,UAAU,CAAC,IAAI,MAAM,CAAC,KAAK,aAAa,GAAG,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC;SACtD,UAAU,CAAC,iBAAiB,EAAE,EAAE,CAAC;SACjC,UAAU,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC;AACzC,CAAC;AAED,qEAAqE;AAErE,MAAM,YAAY,GAAG,gBAAgB,CAAC;AACtC,MAAM,aAAa,GAAG,WAAW,CAAC;AAElC,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;AAC5D,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,YAAY,CAAC,SAAS,GAAG,CAAC,CAAC;IAC3B,OAAO,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC;AACnE,CAAC;AAED,qEAAqE;AAErE,MAAM,kBAAkB,GACtB,4GAA4G,CAAC;AAE/G;;;;;GAKG;AACH,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,kBAAkB,CAAC,SAAS,GAAG,CAAC,CAAC;IACjC,OAAO,IAAI,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC;AAC9C,CAAC"}
+{"version":3,"file":"semantic_tags.js","sourceRoot":"","sources":["../../src/primitives/semantic_tags.ts"],"names":[],"mappings":"AAAA,sDAAsD;AACtD,EAAE;AACF,mFAAmF;AAEnF,MAAM,CAAC,MAAM,aAAa,GAAG,WAAW,CAAC;AAEzC,qEAAqE;AAErE,MAAM,iBAAiB,GAAG,aAAa,CAAC;AACxC,MAAM,iBAAiB,GAAG,qBAAqB,CAAC;AAChD,MAAM,kBAAkB,GAAG,QAAQ,CAAC;AAEpC,MAAM,UAAU,iBAAiB,CAAC,IAAY;IAC5C,OAAO,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AACtC,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,IAAY;IAC9C,kDAAkD;IAClD,iBAAiB,CAAC,SAAS,GAAG,CAAC,CAAC;IAChC,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC;AAC7C,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,OAAO,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;AACvD,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC;AAC7E,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,OAAO,CACL,IAAI,CAAC,QAAQ,CAAC,IAAI,aAAa,GAAG,CAAC;QACnC,IAAI,CAAC,QAAQ,CAAC,IAAI,aAAa,GAAG,CAAC,IAAI,qDAAqD;QAC5F,IAAI,CAAC,QAAQ,CAAC,KAAK,aAAa,GAAG,CAAC;QACpC,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC;QAC/B,IAAI,CAAC,QAAQ,CAAC,iBAAiB,CAAC,CACjC,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,IAAI;SACR,UAAU,CAAC,IAAI,MAAM,CAAC,IAAI,aAAa,WAAW,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC;SAC7D,UAAU,CAAC,IAAI,MAAM,CAAC,KAAK,aAAa,GAAG,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC;SACtD,UAAU,CAAC,iBAAiB,EAAE,EAAE,CAAC;SACjC,UAAU,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC;AACzC,CAAC;AAED,qEAAqE;AAErE,MAAM,YAAY,GAAG,gBAAgB,CAAC;AACtC,MAAM,aAAa,GAAG,WAAW,CAAC;AAElC,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;AAC5D,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,YAAY,CAAC,SAAS,GAAG,CAAC,CAAC;IAC3B,OAAO,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC;AACnE,CAAC;AAED,qEAAqE;AAErE,6FAA6F;AAC7F,qFAAqF;AACrF,2FAA2F;AAC3F,kFAAkF;AAClF,iFAAiF;AACjF,MAAM,kBAAkB,GACtB,oFAAoF,CAAC;AAEvF;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,IAAI,OAAO,GAAG,IAAI,CAAC;IACnB,IAAI,QAAgB,CAAC;IACrB,GAAG,CAAC;QACF,QAAQ,GAAG,OAAO,CAAC;QACnB,kBAAkB,CAAC,SAAS,GAAG,CAAC,CAAC;QACjC,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC,CAAC;IACpD,CAAC,QAAQ,OAAO,KAAK,QAAQ,EAAE;IAC/B,OAAO,OAAO,CAAC;AACjB,CAAC"}

package/dist/primitives/serialize_html.d.ts

@@ -0,0 +1,37 @@
+import type { DocumentViewNode } from './document_view.js';
+import type { Footnote } from './footnotes.js';
+/**
+ * Tracks which footnote numbers were actually rendered as markers in the body, and how many
+ * times each appeared. Used to (a) give repeated markers unique element ids and (b) emit a
+ * back-link from a definition only when its marker was really rendered (no dangling `#fnref-n`).
+ */
+type FootnoteRefs = Map<number, number>;
+export declare function isSafeHref(href: string): boolean;
+/**
+ * Convert one TOON inline-tag string (a `DocumentViewNode.tagged_text` value) to inline HTML.
+ * This is the HTML parallel of `inlineTagsToMarkdown` and the reusable core of the serializer.
+ *
+ * Tag mapping:
+ * - `<b>`/`<i>`/`<u>`/`<a href="...">` → passed through verbatim (already valid, attribute-escaped HTML)
+ * - `<highlight>`                       → `<mark>`
+ * - `<font color=… size=… face=…>`      → `<span style="…">` (values sanitized)
+ *
+ * Literal text spans are HTML-escaped; injected `[^n]` footnote markers become `<sup>` anchors.
+ */
+export declare function inlineTagsToHtml(text: string, refs?: FootnoteRefs): string;
+export interface SerializeHtmlOptions {
+    /** Emit only the body-level elements (no `<!DOCTYPE>`/`<head>`/`<body>` wrapper). */
+    readonly fragment?: boolean;
+    /** Override the `<title>` (full-document mode only). Defaults to the first heading's text. */
+    readonly title?: string;
+}
+/**
+ * Serialize a structured document view to semantic HTML.
+ *
+ * @param nodes     Block nodes from `buildDocumentView({ showFormatting: true }).nodes`.
+ * @param footnotes Footnotes from `DocxDocument.getFootnotes()` (already sorted by
+ *                  `displayNumber`); rendered as a footnotes `<section>` with back-links.
+ */
+export declare function serializeToHtml(nodes: DocumentViewNode[], footnotes?: Footnote[], opts?: SerializeHtmlOptions): string;
+export {};
+//# sourceMappingURL=serialize_html.d.ts.map

package/dist/primitives/serialize_html.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"serialize_html.d.ts","sourceRoot":"","sources":["../../src/primitives/serialize_html.ts"],"names":[],"mappings":"AAgBA,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,oBAAoB,CAAC;AAE3D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAK/C;;;;GAIG;AACH,KAAK,YAAY,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;AA0CxC,wBAAgB,UAAU,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAKhD;AAuCD;;;;;;;;;;GAUG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,YAAY,GAAG,MAAM,CAwB1E;AAiLD,MAAM,WAAW,oBAAoB;IACnC,qFAAqF;IACrF,QAAQ,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;IAC5B,8FAA8F;IAC9F,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC;CACzB;AAED;;;;;;GAMG;AACH,wBAAgB,eAAe,CAC7B,KAAK,EAAE,gBAAgB,EAAE,EACzB,SAAS,GAAE,QAAQ,EAAO,EAC1B,IAAI,GAAE,oBAAyB,GAC9B,MAAM,CA4ER"}