@stll/folio-core 0.32.2 → 0.33.1

package/dist/docx/serializer/paragraphSerializer.js

@@ -210,10 +210,8 @@ function serializeParagraphPropertyChange(change) {
 	const normalizedPreviousPPr = previousPPrInner.length > 0 ? `<w:pPr>${previousPPrInner}</w:pPr>` : "<w:pPr/>";
 	return `<w:pPrChange ${attrs.join(" ")}>${normalizedPreviousPPr}</w:pPrChange>`;
 }
-/**
-* Serialize a hyperlink (w:hyperlink)
-*/
-function serializeHyperlink(hyperlink) {
+/** The attribute list of a `w:hyperlink`, without its children. */
+function hyperlinkAttributes(hyperlink) {
 	const attrs = [];
 	if (hyperlink.rId) attrs.push(`r:id="${escapeXml(hyperlink.rId)}"`);
 	if (hyperlink.anchor) attrs.push(`w:anchor="${escapeXml(hyperlink.anchor)}"`);
@@ -222,12 +220,19 @@ function serializeHyperlink(hyperlink) {
 	if (hyperlink.history === true) attrs.push("w:history=\"1\"");
 	else if (hyperlink.history === false) attrs.push("w:history=\"0\"");
 	if (hyperlink.docLocation) attrs.push(`w:docLocation="${escapeXml(hyperlink.docLocation)}"`);
-	const childrenXml = hyperlink.children.map((child) => {
-		if (child.type === "run") return serializeRun(child);
-		if (child.type === "bookmarkStart") return serializeBookmarkStart(child);
-		return serializeBookmarkEnd(child);
-	}).join("");
-	return `<w:hyperlink${attrs.length > 0 ? ` ${attrs.join(" ")}` : ""}>${childrenXml}</w:hyperlink>`;
+	return attrs.length > 0 ? ` ${attrs.join(" ")}` : "";
+}
+/** One `w:hyperlink` child, with the caller deciding how a run is written. */
+function serializeHyperlinkChild(child, serializeChildRun) {
+	if (child.type === "run") return serializeChildRun(child);
+	return child.type === "bookmarkStart" ? serializeBookmarkStart(child) : serializeBookmarkEnd(child);
+}
+/**
+* Serialize a hyperlink (w:hyperlink)
+*/
+function serializeHyperlink(hyperlink) {
+	const childrenXml = hyperlink.children.map((child) => serializeHyperlinkChild(child, serializeRun)).join("");
+	return `<w:hyperlink${hyperlinkAttributes(hyperlink)}>${childrenXml}</w:hyperlink>`;
 }
 /**
 * Serialize bookmark start (w:bookmarkStart)
@@ -368,6 +373,14 @@ function serializeMoveRangeStart(tag, marker) {
 function rewriteRunTextAsDeleted(xml) {
 	return xml.replace(/<w:t\b/gu, "<w:delText").replace(/<\/w:t>/gu, "</w:delText>").replace(/<w:instrText\b/gu, "<w:delInstrText").replace(/<\/w:instrText>/gu, "</w:delInstrText>");
 }
+function trackedChangeTag(change) {
+	switch (change.type) {
+		case "insertion": return "ins";
+		case "deletion": return "del";
+		case "moveFrom": return "moveFrom";
+		case "moveTo": return "moveTo";
+	}
+}
 function serializeTrackedChange(tag, change) {
 	const info = change.info;
 	const normalizedId = normalizeRevisionId(info.id);
@@ -389,14 +402,39 @@ function serializeTrackedChange(tag, change) {
 			return rewriteRunTextAsDeleted(contentXml);
 		}).join("");
 	};
-	const contentXml = change.content.map((item) => {
-		if (item.type === "run") {
-			if (tag === "del" || tag === "moveFrom") return serializeDeletedRun(item);
-			return serializeRun(item);
+	const serializeContentRun = (run) => tag === "del" || tag === "moveFrom" ? serializeDeletedRun(run) : serializeRun(run);
+	const serializeWrappedItem = (item) => {
+		if (item.type === "run") return serializeContentRun(item);
+		if (item.type === "simpleField" || item.type === "complexField") {
+			const xml = item.type === "simpleField" ? serializeSimpleField(item) : serializeComplexField(item);
+			return tag === "del" || tag === "moveFrom" ? rewriteRunTextAsDeleted(xml) : xml;
 		}
-		return serializeHyperlink(item);
-	}).join("");
-	return `<w:${tag} ${attrs.join(" ")}>${contentXml}</w:${tag}>`;
+		if (item.type === "insertion" || item.type === "deletion" || item.type === "moveFrom" || item.type === "moveTo") return serializeTrackedChange(trackedChangeTag(item), item);
+		return item.type === "bookmarkStart" ? serializeBookmarkStart(item) : serializeBookmarkEnd(item);
+	};
+	const open = `<w:${tag} ${attrs.join(" ")}>`;
+	const close = `</w:${tag}>`;
+	const wrap = (inner) => inner.length === 0 ? "" : `${open}${inner}${close}`;
+	if (change.content.length === 0) return `${open}${close}`;
+	const segments = [];
+	const pending = [];
+	const flushPending = () => {
+		if (pending.length > 0) {
+			segments.push(wrap(pending.join("")));
+			pending.length = 0;
+		}
+	};
+	for (const item of change.content) {
+		if (item.type === "hyperlink") {
+			flushPending();
+			const childrenXml = item.children.map((child) => serializeHyperlinkChild(child, serializeContentRun)).join("");
+			segments.push(`<w:hyperlink${hyperlinkAttributes(item)}>${open}${childrenXml}${close}</w:hyperlink>`);
+			continue;
+		}
+		pending.push(serializeWrappedItem(item));
+	}
+	flushPending();
+	return segments.join("");
 }
 /** Emit the `<w:commentReference>` run Word places after a comment range end. */
 function serializeCommentReferenceRun(id) {
@@ -470,12 +508,27 @@ function serializeParagraph(paragraph) {
 	return `<w:p${attrsStr}>${parts.join("")}</w:p>`;
 }
 function injectRenderedPageBreakIntoFirstRun(xml) {
-	const runOpeningTag = /<w:r(?=[\s>/])[^>]*>/u;
-	const openingTag = runOpeningTag.exec(xml);
+	const openingTag = /<w:r(?=[\s>/])[^>]*>/u.exec(xml);
 	if (!openingTag) return null;
 	const runEnd = xml.indexOf("</w:r>", openingTag.index + openingTag[0].length);
 	if (runEnd !== -1 && xml.slice(openingTag.index + openingTag[0].length, runEnd).includes("<w:lastRenderedPageBreak/>")) return xml;
-	return xml.replace(runOpeningTag, (match) => `${match}<w:lastRenderedPageBreak/>`);
+	const contentStart = openingTag.index + openingTag[0].length;
+	let insertionOffset = contentStart;
+	if (xml.startsWith("<w:rPr", contentStart)) {
+		const propertyTag = /<\/?w:rPr(?=[\s>/])[^>]*>/gu;
+		propertyTag.lastIndex = contentStart;
+		let depth = 0;
+		for (const match of xml.matchAll(propertyTag)) {
+			if (match.index !== contentStart && depth === 0) break;
+			if (match[0].startsWith("</")) depth--;
+			else if (!match[0].endsWith("/>")) depth++;
+			if (depth === 0) {
+				insertionOffset = match.index + match[0].length;
+				break;
+			}
+		}
+	}
+	return `${xml.slice(0, insertionOffset)}<w:lastRenderedPageBreak/>${xml.slice(insertionOffset)}`;
 }
 /**
 * Serialize multiple paragraphs to OOXML XML

package/dist/docx/serializer/runSerializer.js

@@ -92,13 +92,6 @@ function serializeTextFormatting(formatting) {
 		if (formatting.fontFamily.csTheme) fontAttrs.push(`w:cstheme="${escapeXml(formatting.fontFamily.csTheme)}"`);
 		if (fontAttrs.length > 0) parts.push(`<w:rFonts ${fontAttrs.join(" ")}/>`);
 	}
-	if (formatting.language) {
-		const languageAttrs = [];
-		if (formatting.language.val) languageAttrs.push(`w:val="${escapeXml(formatting.language.val)}"`);
-		if (formatting.language.eastAsia) languageAttrs.push(`w:eastAsia="${escapeXml(formatting.language.eastAsia)}"`);
-		if (formatting.language.bidi) languageAttrs.push(`w:bidi="${escapeXml(formatting.language.bidi)}"`);
-		if (languageAttrs.length > 0) parts.push(`<w:lang ${languageAttrs.join(" ")}/>`);
-	}
 	if (formatting.bold === true) parts.push("<w:b/>");
 	else if (formatting.bold === false) parts.push("<w:b w:val=\"0\"/>");
 	if (formatting.boldCs === true) parts.push("<w:bCs/>");
@@ -134,11 +127,12 @@ function serializeTextFormatting(formatting) {
 	if (formatting.position !== void 0) parts.push(`<w:position w:val="${intAttr(formatting.position)}"/>`);
 	if (formatting.fontSize !== void 0) parts.push(`<w:sz w:val="${intAttr(formatting.fontSize)}"/>`);
 	if (formatting.fontSizeCs !== void 0) parts.push(`<w:szCs w:val="${intAttr(formatting.fontSizeCs)}"/>`);
+	let customHighlightShadingXml = "";
 	if (formatting.highlight) {
 		if (VALID_HIGHLIGHT_COLORS.has(formatting.highlight)) parts.push(`<w:highlight w:val="${formatting.highlight}"/>`);
 		else if (!formatting.shading) {
 			const hex = formatting.highlight.replace(/^#/u, "");
-			if (/^[0-9a-fA-F]{6}$/u.test(hex)) parts.push(`<w:shd w:val="clear" w:color="auto" w:fill="${hex}"/>`);
+			if (/^[0-9a-fA-F]{6}$/u.test(hex)) customHighlightShadingXml = `<w:shd w:val="clear" w:color="auto" w:fill="${hex}"/>`;
 		}
 	}
 	if (formatting.underline) {
@@ -152,14 +146,21 @@ function serializeTextFormatting(formatting) {
 		parts.push(`<w:u ${uAttrs.join(" ")}/>`);
 	}
 	if (formatting.effect && formatting.effect !== "none") parts.push(`<w:effect w:val="${formatting.effect}"/>`);
-	if (formatting.emphasisMark) parts.push(`<w:em w:val="${formatting.emphasisMark}"/>`);
-	const shadingXml = serializeShading(formatting.shading);
+	const shadingXml = serializeShading(formatting.shading) || customHighlightShadingXml;
 	if (shadingXml) parts.push(shadingXml);
 	if (formatting.vertAlign) parts.push(`<w:vertAlign w:val="${formatting.vertAlign}"/>`);
 	if (formatting.rtl === true) parts.push("<w:rtl/>");
 	else if (formatting.rtl === false) parts.push("<w:rtl w:val=\"0\"/>");
 	if (formatting.cs === true) parts.push("<w:cs/>");
 	else if (formatting.cs === false) parts.push("<w:cs w:val=\"0\"/>");
+	if (formatting.emphasisMark) parts.push(`<w:em w:val="${formatting.emphasisMark}"/>`);
+	if (formatting.language) {
+		const languageAttrs = [];
+		if (formatting.language.val) languageAttrs.push(`w:val="${escapeXml(formatting.language.val)}"`);
+		if (formatting.language.eastAsia) languageAttrs.push(`w:eastAsia="${escapeXml(formatting.language.eastAsia)}"`);
+		if (formatting.language.bidi) languageAttrs.push(`w:bidi="${escapeXml(formatting.language.bidi)}"`);
+		if (languageAttrs.length > 0) parts.push(`<w:lang ${languageAttrs.join(" ")}/>`);
+	}
 	if (parts.length === 0) return "";
 	return `<w:rPr>${parts.join("")}</w:rPr>`;
 }

package/dist/docx/serializer/tableSerializer.js

@@ -250,7 +250,8 @@ function serializeTableCellFormatting(formatting, propertyChanges, structuralCha
 		if (formatting.textDirection) parts.push(`<w:textDirection w:val="${formatting.textDirection}"/>`);
 		if (formatting.fitText) parts.push("<w:tcFitText/>");
 		if (formatting.verticalAlign) parts.push(`<w:vAlign w:val="${formatting.verticalAlign}"/>`);
-		if (formatting.hideMark !== void 0) parts.push(formatting.hideMark ? "<w:hideMark/>" : "<w:hideMark w:val=\"false\"/>");
+		if (formatting.hideMark === true) parts.push("<w:hideMark/>");
+		else if (formatting.hideMark === false) parts.push("<w:hideMark w:val=\"off\"/>");
 	}
 	if (structuralChange) {
 		if (structuralChange.type === "tableCellInsertion") parts.push(`<w:cellIns ${serializeTrackedChangeAttributes(structuralChange.info)}/>`);
@@ -276,11 +277,30 @@ function serializeTableCellPropertyChange(change) {
 	return `<w:tcPrChange ${attrs}>${previousTcPrInner.length > 0 ? `<w:tcPr>${previousTcPrInner}</w:tcPr>` : "<w:tcPr/>"}</w:tcPrChange>`;
 }
 /**
-* Serialize table grid (w:tblGrid)
+* Columns the grid has to declare: the widest row's span total, because a grid
+* narrower than a row leaves cells with no column to sit in.
 */
-function serializeTableGrid(columnWidths) {
-	if (!columnWidths || columnWidths.length === 0) return "";
-	return `<w:tblGrid>${columnWidths.map((w) => `<w:gridCol w:w="${intAttr(w)}"/>`).join("")}</w:tblGrid>`;
+function gridColumnCount(table) {
+	let widest = 0;
+	for (const row of table.rows) {
+		let columns = 0;
+		for (const cell of row.cells) columns += cell.formatting?.gridSpan ?? 1;
+		widest = Math.max(widest, columns);
+	}
+	return widest;
+}
+/**
+* Serialize table grid (w:tblGrid).
+*
+* `w:tblGrid` is required on every `w:tbl` and `w:w` is optional on a
+* `w:gridCol`, so a table whose column widths were never measured — one the
+* comparison creates, say — still declares a grid, just without widths.
+*/
+function serializeTableGrid(table) {
+	const columnWidths = table.columnWidths;
+	if (columnWidths && columnWidths.length > 0) return `<w:tblGrid>${columnWidths.map((w) => `<w:gridCol w:w="${intAttr(w)}"/>`).join("")}</w:tblGrid>`;
+	const columns = gridColumnCount(table);
+	return columns === 0 ? "<w:tblGrid/>" : `<w:tblGrid>${"<w:gridCol/>".repeat(columns)}</w:tblGrid>`;
 }
 /**
 * Serialize cell content (paragraphs, nested tables)
@@ -319,11 +339,7 @@ function serializeTableRow(row, serializeParagraph) {
 * @returns XML string for the table
 */
 function serializeTable(table, serializeParagraph) {
-	const parts = [];
-	const tblPrXml = serializeTableFormatting(table.formatting, table.propertyChanges);
-	if (tblPrXml) parts.push(tblPrXml);
-	const tblGridXml = serializeTableGrid(table.columnWidths);
-	if (tblGridXml) parts.push(tblGridXml);
+	const parts = [serializeTableFormatting(table.formatting, table.propertyChanges) || "<w:tblPr/>", serializeTableGrid(table)];
 	for (const row of table.rows) parts.push(serializeTableRow(row, serializeParagraph));
 	return `<w:tbl>${parts.join("")}</w:tbl>`;
 }

package/dist/docx/server/applyDocxXmlPatchProposal.js

@@ -2,8 +2,8 @@ import { loadDocxArchive } from "./boundedArchive.js";
 import { FOLIO_DOCX_XML_PATCH_PROPOSAL_PROFILE, InvalidFolioDocxXmlPatchProposalError, evaluateDocxXmlPatchProposal, parseFolioDocxXmlPatchProposal } from "./evaluateDocxXmlPatchProposal.js";
 import { FOLIO_DOCX_CONFORMANCE_PROFILE, validateDocxConformance } from "./validateDocxConformance.js";
 import { TaggedError, panic } from "better-result";
-import JSZip from "jszip";
 import { createHash } from "node:crypto";
+import JSZip from "jszip";
 //#region src/docx/server/applyDocxXmlPatchProposal.ts
 const FOLIO_DOCX_XML_PATCH_APPLICATION_VERSION = 1;
 const FOLIO_DOCX_XML_PATCH_APPLICATION_PROFILE = "folio-xml-patch-application-v1";

package/dist/docx/server/createBilingualDocx.js

@@ -1,4 +1,5 @@
 import { FolioDocxReviewer } from "../../ai-edits/headless.js";
+import { isFolioAIContentBlock } from "../../ai-edits/snapshot.js";
 import { toArrayBuffer } from "../../utils/docxInput.js";
 import { ensureParaIds } from "../ensureParaIds.js";
 import { parseDocx } from "../parser.js";
@@ -14,7 +15,7 @@ import { createBilingualDocument, readBilingualDocument } from "./createBilingua
 */
 async function createBilingualDocx(input, options) {
 	const stampedBuffer = await toArrayBuffer((await ensureParaIds(input)).docx);
-	const editableParagraphIds = new Set((await FolioDocxReviewer.fromBuffer(stampedBuffer)).snapshot().blocks.map(({ id }) => id));
+	const editableParagraphIds = new Set((await FolioDocxReviewer.fromBuffer(stampedBuffer)).snapshot().blocks.filter(isFolioAIContentBlock).map(({ id }) => id));
 	const { document, warnings } = createBilingualDocument(await parseDocx(stampedBuffer, { preloadFonts: false }), {
 		...options,
 		editableParagraphIds
@@ -29,7 +30,7 @@ async function createBilingualDocx(input, options) {
 /** Bytes-in form of {@link readBilingualDocument}. */
 async function readBilingualDocx(input) {
 	const buffer = await toArrayBuffer(input);
-	return readBilingualDocument(await parseDocx(buffer, { preloadFonts: false }), new Set((await FolioDocxReviewer.fromBuffer(buffer)).snapshot().blocks.map(({ id }) => id)));
+	return readBilingualDocument(await parseDocx(buffer, { preloadFonts: false }), new Set((await FolioDocxReviewer.fromBuffer(buffer)).snapshot().blocks.filter(isFolioAIContentBlock).map(({ id }) => id)));
 }
 //#endregion
 export { createBilingualDocx, readBilingualDocx };

package/dist/docx/streamingXmlParser.d.ts

@@ -11,6 +11,18 @@ type ParseXmlResult = {
  * pass. Unsupported or malformed constructs return a sentinel so callers can
  * retain the general-purpose parser as a compatibility fallback.
  */
+type AttributeValueSpan = {
+  start: number;
+  end: number;
+};
+type OpenTagVisitor = (element: XmlElement, attributeValueSpans: ReadonlyMap<string, AttributeValueSpan>) => ReadonlyMap<string, string> | null;
 declare const parseStreamingXml: (xml: string) => ParseXmlResult;
+/** Rewrite selected decimal attribute values while preserving every other source byte. */
+declare const rewriteStreamingXmlDecimalAttributes: (xml: string, visitOpenTag: OpenTagVisitor) => {
+  status: "rewritten";
+  value: string;
+} | {
+  status: "unsupported";
+};
 //#endregion
-export { parseStreamingXml };
+export { parseStreamingXml, rewriteStreamingXmlDecimalAttributes };

package/dist/docx/streamingXmlParser.js

@@ -1,3 +1,4 @@
+import { attachXmlNamespaceContext } from "./xmlParser.js";
 import { FOLIO_XML_RESOURCE_LIMITS } from "./xmlResourceLimits.js";
 //#region src/docx/streamingXmlParser.ts
 const BUILT_IN_ENTITIES = {
@@ -7,14 +8,10 @@ const BUILT_IN_ENTITIES = {
 	lt: "<",
 	quot: "\""
 };
-/**
-* Parse ordinary OOXML into Folio's existing element representation in one
-* pass. Unsupported or malformed constructs return a sentinel so callers can
-* retain the general-purpose parser as a compatibility fallback.
-*/
-const parseStreamingXml = (xml) => {
+const parseStreamingXmlInternal = (xml, visitOpenTag) => {
 	const root = { elements: [] };
 	const stack = [];
+	const replacements = [];
 	let cursor = 0;
 	let mergeAdjacentText = false;
 	while (cursor < xml.length) {
@@ -55,9 +52,19 @@ const parseStreamingXml = (xml) => {
 			mergeAdjacentText = false;
 			continue;
 		}
-		const parsedTag = parseOpenTag(xml, open + 1, close);
+		const parsedTag = parseOpenTag(xml, open + 1, close, visitOpenTag !== void 0);
 		if (parsedTag.status === "unsupported") return parsedTag;
 		const parent = stack.at(-1)?.element ?? root;
+		attachXmlNamespaceContext(parsedTag.element, parent.namespaceScope);
+		const rewritten = visitOpenTag?.(parsedTag.element, parsedTag.attributeValueSpans ?? /* @__PURE__ */ new Map());
+		if (rewritten) for (const [attributeName, value] of rewritten) {
+			const span = parsedTag.attributeValueSpans?.get(attributeName);
+			if (!span) return { status: "unsupported" };
+			replacements.push({
+				...span,
+				value
+			});
+		}
 		appendElement(parent, parsedTag.element);
 		if (!parsedTag.selfClosing) {
 			if (stack.length >= FOLIO_XML_RESOURCE_LIMITS.maxDepth) return { status: "unsupported" };
@@ -72,10 +79,35 @@ const parseStreamingXml = (xml) => {
 	if (stack.length > 0) return { status: "unsupported" };
 	return {
 		status: "parsed",
-		value: root
+		value: root,
+		replacements
+	};
+};
+const parseStreamingXml = (xml) => {
+	const parsed = parseStreamingXmlInternal(xml);
+	return parsed.status === "parsed" ? {
+		status: "parsed",
+		value: parsed.value
+	} : { status: "unsupported" };
+};
+/** Rewrite selected decimal attribute values while preserving every other source byte. */
+const rewriteStreamingXmlDecimalAttributes = (xml, visitOpenTag) => {
+	const parsed = parseStreamingXmlInternal(xml, visitOpenTag);
+	if (parsed.status === "unsupported") return parsed;
+	const chunks = [];
+	let cursor = 0;
+	for (const replacement of parsed.replacements.toSorted((left, right) => left.start - right.start)) {
+		if (!/^\d+$/u.test(replacement.value)) return { status: "unsupported" };
+		chunks.push(xml.slice(cursor, replacement.start), replacement.value);
+		cursor = replacement.end;
+	}
+	chunks.push(xml.slice(cursor));
+	return {
+		status: "rewritten",
+		value: chunks.join("")
 	};
 };
-const parseOpenTag = (xml, start, close) => {
+const parseOpenTag = (xml, start, close, captureAttributeSpans) => {
 	let cursor = skipWhitespace(xml, start, close);
 	const nameStart = cursor;
 	cursor = scanName(xml, cursor, close);
@@ -83,6 +115,7 @@ const parseOpenTag = (xml, start, close) => {
 	const name = xml.slice(nameStart, cursor);
 	if (isUnsafePropertyName(name)) return { status: "unsupported" };
 	let attributes;
+	const attributeValueSpans = captureAttributeSpans ? /* @__PURE__ */ new Map() : void 0;
 	let selfClosing = false;
 	while (cursor < close) {
 		cursor = skipWhitespace(xml, cursor, close);
@@ -110,6 +143,10 @@ const parseOpenTag = (xml, start, close) => {
 		if (decoded === null) return { status: "unsupported" };
 		attributes ??= {};
 		attributes[attributeName] = decoded;
+		attributeValueSpans?.set(attributeName, {
+			start: valueStart,
+			end: cursor
+		});
 		cursor += 1;
 	}
 	const element = {
@@ -121,7 +158,8 @@ const parseOpenTag = (xml, start, close) => {
 		status: "parsed",
 		element,
 		name,
-		selfClosing
+		selfClosing,
+		attributeValueSpans
 	};
 };
 const appendElement = (parent, child) => {
@@ -246,4 +284,4 @@ const isUnsafePropertyName = (name) => {
 	}
 };
 //#endregion
-export { parseStreamingXml };
+export { parseStreamingXml, rewriteStreamingXmlDecimalAttributes };

package/dist/docx/unzip.d.ts

@@ -59,7 +59,6 @@ type RawDocxContent = {
  * @returns Promise resolving to extracted content
  */
 declare function unzipDocx(buffer: ArrayBuffer, options?: DocxUnzipOptions): Promise<RawDocxContent>;
-declare function isPreservableDocxEntry(path: string): boolean;
 /**
  * Get a list of all files in the DOCX
  *
@@ -120,4 +119,4 @@ declare function getContentSummary(content: RawDocxContent): {
   totalFiles: number;
 };
 //#endregion
-export { DocxSecurityError, DocxUnzipLimits, DocxUnzipOptions, RawDocxContent, extractFile, getContentSummary, getFileList, getMediaMimeType, hasFile, isPreservableDocxEntry, mediaToDataUrl, unzipDocx };
+export { DocxSecurityError, DocxUnzipLimits, DocxUnzipOptions, RawDocxContent, extractFile, getContentSummary, getFileList, getMediaMimeType, hasFile, mediaToDataUrl, unzipDocx };

package/dist/docx/unzip.js

@@ -137,13 +137,13 @@ async function unzipDocx(buffer, options = {}) {
 	const extractionTasks = [];
 	for (const [path, file] of entries) {
 		if (!isSafeDocxPath(path)) throw new DocxSecurityError("DOCX file contains an unsafe entry path");
-		if (!isPreservableDocxEntry(path)) continue;
-		const lowerPath = path.toLowerCase();
 		const declaredSize = getEntryUncompressedSize(file);
 		if (declaredSize !== null) {
 			totalUncompressedBytes += declaredSize;
 			if (totalUncompressedBytes > limits.maxTotalUncompressedBytes) throw new DocxSecurityError("DOCX file expands beyond the maximum allowed size");
 		}
+		if (!isParsedDocxEntry(path)) continue;
+		const lowerPath = path.toLowerCase();
 		if (lowerPath.endsWith(".xml") || lowerPath.endsWith(".rels")) {
 			assertEntrySize(path, declaredSize, limits.maxXmlBytes);
 			if (options.extractAllXml === false && !shouldExtractXmlPart(lowerPath)) continue;
@@ -328,9 +328,28 @@ function isSafeDocxPath(path) {
 	if (!path || path.startsWith("/") || path.includes("\\")) return false;
 	return !path.split("/").some((part) => part === "..");
 }
-function isPreservableDocxEntry(path) {
+/**
+* Parts folio never looks inside. A macro project and an ActiveX control carry
+* code; folio models neither, so it reads neither — not into the document
+* model, not even into the raw XML map. The save path carries them anyway,
+* straight from the source archive, which is what keeps a `.docm` a `.docm`.
+*/
+const UNINTERPRETED_PART_NAMES = /* @__PURE__ */ new Set(["word/vbaproject.bin", "word/vbadata.xml"]);
+const isUninterpretedPart = (lowerPath) => UNINTERPRETED_PART_NAMES.has(lowerPath) || lowerPath.startsWith("word/activex/");
+/**
+* Whether this entry is one the PARSER reads into the document model. It says
+* nothing about what a save keeps: a repack hands back every part of the
+* package (see `packageParts.ts`), including the ones this returns false for.
+*
+* That asymmetry is the point. A macro project, an ActiveX control and an OLE
+* embedding are opaque bytes to folio — never decoded, never interpreted,
+* never executed — and they travel from the source archive to the saved one
+* without any code here looking inside them.
+*/
+function isParsedDocxEntry(path) {
 	if (!isSafeDocxPath(path)) return false;
 	const lowerPath = path.toLowerCase();
+	if (isUninterpretedPart(lowerPath)) return false;
 	if (lowerPath.startsWith("word/media/")) return PRESERVABLE_MEDIA_MIME_TYPES.has(getMediaMimeType(path));
 	if (lowerPath.startsWith("docprops/thumbnail.")) return PRESERVABLE_MEDIA_MIME_TYPES.has(getMediaMimeType(path));
 	if (lowerPath.startsWith("word/fonts/")) return true;
@@ -374,7 +393,7 @@ function isMediaContentAllowed(data, mimeType) {
 */
 function getFileList(content) {
 	const files = [];
-	for (const path of Object.keys(content.originalZip.files)) if (!content.originalZip.files[path]?.dir && isPreservableDocxEntry(path)) files.push(path);
+	for (const path of Object.keys(content.originalZip.files)) if (!content.originalZip.files[path]?.dir && isParsedDocxEntry(path)) files.push(path);
 	return files.toSorted();
 }
 /**
@@ -425,7 +444,7 @@ function mediaToDataUrl(data, mimeType) {
 */
 function extractFile(content, path) {
 	const file = content.originalZip.file(path);
-	if (!file || !isPreservableDocxEntry(path)) return Promise.resolve(null);
+	if (!file || !isParsedDocxEntry(path)) return Promise.resolve(null);
 	const lowerPath = path.toLowerCase();
 	if (lowerPath.endsWith(".xml") || lowerPath.endsWith(".rels")) return file.async("text");
 	return file.async("arraybuffer");
@@ -464,4 +483,4 @@ function getContentSummary(content) {
 	};
 }
 //#endregion
-export { DocxSecurityError, extractFile, getContentSummary, getFileList, getMediaMimeType, hasFile, isPreservableDocxEntry, mediaToDataUrl, unzipDocx };
+export { DocxSecurityError, extractFile, getContentSummary, getFileList, getMediaMimeType, hasFile, mediaToDataUrl, unzipDocx };

package/dist/docx/xmlParser.d.ts

@@ -47,6 +47,8 @@ type XmlNamespaceScope = {
   bindings: ReadonlyMap<string, string>;
   parent?: XmlNamespaceScope;
 };
+/** Attach the element's resolved namespace metadata from its in-scope declarations. */
+declare const attachXmlNamespaceContext: (element: XmlElement, inheritedNamespaceScope?: XmlNamespaceScope) => XmlNamespaceScope;
 /**
  * Common OOXML namespace URIs — re-exported from @stll/docx-utils.
  */
@@ -344,4 +346,4 @@ declare function mergeXmlnsDeclarations(inherited: Record<string, string>, eleme
  */
 declare function cloneWithXmlnsDeclarations(element: XmlElement, xmlnsDecls: Record<string, string>): XmlElement;
 //#endregion
-export { NAMESPACES, WORDPROCESSINGML_NAMESPACE_URIS, XmlAttributeMatch, XmlElement, XmlNamespaceScope, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, hasFlag, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument };
+export { NAMESPACES, WORDPROCESSINGML_NAMESPACE_URIS, XmlAttributeMatch, XmlElement, XmlNamespaceScope, attachXmlNamespaceContext, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, hasFlag, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument };

package/dist/docx/xmlParser.js

@@ -64,6 +64,37 @@ const resolveNamespaceUri = (scope, prefix) => {
 		current = current.parent;
 	}
 };
+/** Attach the element's resolved namespace metadata from its in-scope declarations. */
+const attachXmlNamespaceContext = (element, inheritedNamespaceScope = EMPTY_NAMESPACE_SCOPE) => {
+	let localBindings = null;
+	if (element.attributes) for (const [attribute, value] of Object.entries(element.attributes)) {
+		if (typeof value !== "string" || attribute !== "xmlns" && !attribute.startsWith("xmlns:")) continue;
+		localBindings ??= /* @__PURE__ */ new Map();
+		const prefix = attribute === "xmlns" ? "" : attribute.slice(6);
+		localBindings.set(prefix, value);
+	}
+	const namespaceScope = localBindings === null ? inheritedNamespaceScope : {
+		bindings: localBindings,
+		parent: inheritedNamespaceScope
+	};
+	Object.defineProperty(element, "namespaceScope", {
+		configurable: false,
+		enumerable: false,
+		value: namespaceScope,
+		writable: false
+	});
+	const name = element.name ?? "";
+	const colonIndex = name.indexOf(":");
+	const prefix = colonIndex === -1 ? "" : name.slice(0, colonIndex);
+	const namespaceUri = resolveNamespaceUri(namespaceScope, prefix);
+	if (namespaceUri !== void 0) Object.defineProperty(element, "namespaceUri", {
+		configurable: false,
+		enumerable: false,
+		value: namespaceUri,
+		writable: false
+	});
+	return namespaceScope;
+};
 /**
 * Convert a fast-xml-parser preserveOrder node into an XmlElement.
 *
@@ -85,32 +116,7 @@ function fxpNodeToElement(node, inheritedNamespaceScope = EMPTY_NAMESPACE_SCOPE)
 			name: key
 		};
 		if (attrs) element.attributes = attrs;
-		let localBindings = null;
-		if (attrs) for (const [attribute, value] of Object.entries(attrs)) {
-			if (attribute !== "xmlns" && !attribute.startsWith("xmlns:")) continue;
-			localBindings ??= /* @__PURE__ */ new Map();
-			const prefix = attribute === "xmlns" ? "" : attribute.slice(6);
-			localBindings.set(prefix, value);
-		}
-		const namespaceScope = localBindings === null ? inheritedNamespaceScope : {
-			bindings: localBindings,
-			parent: inheritedNamespaceScope
-		};
-		Object.defineProperty(element, "namespaceScope", {
-			configurable: false,
-			enumerable: false,
-			value: namespaceScope,
-			writable: false
-		});
-		const colonIndex = key.indexOf(":");
-		const prefix = colonIndex === -1 ? "" : key.slice(0, colonIndex);
-		const namespaceUri = resolveNamespaceUri(namespaceScope, prefix);
-		if (namespaceUri !== void 0) Object.defineProperty(element, "namespaceUri", {
-			configurable: false,
-			enumerable: false,
-			value: namespaceUri,
-			writable: false
-		});
+		const namespaceScope = attachXmlNamespaceContext(element, inheritedNamespaceScope);
 		if (children.length > 0) {
 			for (let index = 0; index < children.length; index += 1) children[index] = fxpNodeToElement(children[index], namespaceScope);
 			element.elements = children;
@@ -745,4 +751,4 @@ function cloneWithXmlnsDeclarations(element, xmlnsDecls) {
 	return element;
 }
 //#endregion
-export { NAMESPACES, WORDPROCESSINGML_NAMESPACE_URIS, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, hasFlag, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument };
+export { NAMESPACES, WORDPROCESSINGML_NAMESPACE_URIS, attachXmlNamespaceContext, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, hasFlag, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument };

package/dist/export-pdf.d.ts

@@ -0,0 +1,43 @@
+import { DocxInput } from "./utils/docxInput.js";
+import { DisplayMetadata, DisplayUnsupported } from "./display-list/types.js";
+import { HeadlessFontSource, HeadlessFontSubstitution } from "./fonts/headlessMeasure.js";
+import { HeadlessLayoutGap } from "./headless-layout.js";
+import { PdfSubstitution } from "./pdf/fonts.js";
+import { PdfUnencodable } from "./pdf/writePdf.js";
+import { Result } from "better-result";
+//#region src/export-pdf.d.ts
+declare const ExportPdfError_base: import("better-result").TaggedErrorClass<"ExportPdfError">;
+declare class ExportPdfError extends ExportPdfError_base<{
+  message: string;
+  cause?: unknown;
+}> {}
+type ExportDocxToPdfOptions = {
+  /** Supplies face binaries for measurement and for embedding alike. */
+  readonly fonts: HeadlessFontSource;
+  /**
+   * ISO-8601 instant stamped as the PDF's creation and modification date.
+   * Required rather than defaulted so a caller cannot get a nondeterministic
+   * document by omission; pass the source package's own timestamp or a fixed
+   * epoch.
+   */
+  readonly timestamp: string;
+  readonly metadata?: DisplayMetadata;
+  readonly producer?: string;
+};
+type ExportDocxToPdfResult = {
+  readonly bytes: Uint8Array;
+  readonly pageCount: number;
+  /** Constructs the display list could not represent. */
+  readonly unsupported: readonly DisplayUnsupported[];
+  /** Stories the headless pipeline does not paginate. */
+  readonly layoutGaps: readonly HeadlessLayoutGap[];
+  /** Faces measured with a stand-in. */
+  readonly measurementSubstitutions: readonly HeadlessFontSubstitution[];
+  /** Faces embedded as a stand-in. */
+  readonly embeddingSubstitutions: readonly PdfSubstitution[];
+  /** Code points painted as `.notdef` because no supplied face covers them. */
+  readonly unencodable: readonly PdfUnencodable[];
+};
+declare const exportDocxToPdf: (input: DocxInput, options: ExportDocxToPdfOptions) => Promise<Result<ExportDocxToPdfResult, ExportPdfError>>;
+//#endregion
+export { ExportDocxToPdfOptions, ExportDocxToPdfResult, ExportPdfError, exportDocxToPdf };