@stll/folio-core 0.45.0 → 0.46.0

package/dist/docx/tableParser.js

@@ -1,7 +1,8 @@
-import { attachPendingRangeMarkers, attachTrailingRangeMarkers, isBlockRangeMarker } from "./blockRangeMarkers.js";
+import { NO_MODELLED_ATTRIBUTES, attributeRemainder } from "./attributeRemainder.js";
 import { parseBookmarkEnd, parseBookmarkStart } from "./bookmarkParser.js";
 import { appendBookmarkMarkerToLastParagraphInBlocks, appendBookmarkMarkerToLastParagraphInCells, prependBookmarkMarkersToFirstParagraphInBlocks, prependBookmarkMarkersToFirstParagraphInCell } from "./bookmarkPlacement.js";
 import { parseBorderSpec } from "./borderParser.js";
+import { CAPTURE, OWNED_ELSEWHERE, dispatchChildren, withPreservedChildren } from "./containerChildren.js";
 import { parseParagraph } from "./paragraphParser.js";
 import { enrichParagraphTextBoxes } from "./paragraphTextBoxEnrichment.js";
 import { FloatingTableXSpecSchema, FloatingTableYSpecSchema, TableCellTextDirectionSchema, narrowEnum } from "./parserEnums.js";
@@ -10,7 +11,7 @@ import { TABLE_LOOK_FLAGS } from "./tableLook.js";
 import { parsePropertyChangeInfo, parseTrackedChangeInfo } from "./trackedChangeInfo.js";
 import { percentageSpelling, transitionalSlotEncoding } from "./transitionalSpelling.js";
 import { captureVerbatimXml } from "./verbatimCapture.js";
-import { cloneElement, findChild, findChildByLocalName, findChildren, getAttribute, getChildElements, getLocalName, mergeXmlnsDeclarations, parseBooleanElement, parseNumericAttribute, parseOnOffAttribute, parseTableMeasurementValue, selectAlternateContentBranch } from "./xmlParser.js";
+import { cloneElement, findChild, findChildByLocalName, findChildren, getAttribute, getLocalName, mergeXmlnsDeclarations, parseBooleanElement, parseNumericAttribute, parseOnOffAttribute, parseTableMeasurementValue, selectAlternateContentBranch } from "./xmlParser.js";
 //#region src/docx/tableParser.ts
 /**
 * Sanity cap on `w:gridSpan` (and the derived table column count). Word's
@@ -472,59 +473,87 @@ function withContainerXmlns(options, element) {
 	};
 }
 function parseCellContent(tcElement, styles, theme, numbering, rels, media, options) {
-	const content = [];
+	const modelled = [];
 	const pendingBookmarkMarkers = [];
-	const pendingRangeMarkers = [];
-	const elements = getChildElements(tcElement);
-	const parseCellChild = (child, childOptions = options) => {
-		if (!child.name) return;
-		const localName = getLocalName(child.name);
-		if (localName === "p") {
-			const para = parseParagraph(child, styles, theme, numbering, rels, media, childOptions);
-			enrichParagraphTextBoxes(para, child, styles, theme, numbering, rels, media, parseTable);
-			prependPendingBookmarkMarkers(para, pendingBookmarkMarkers);
-			attachPendingRangeMarkers(para, pendingRangeMarkers);
-			content.push(para);
-			return;
-		}
-		if (localName === "tbl") {
-			const table = parseTable(child, styles, theme, numbering, rels, media, childOptions);
-			if (!table) return;
-			if (prependBookmarkMarkersToFirstParagraphInBlocks([table], pendingBookmarkMarkers)) pendingBookmarkMarkers.length = 0;
-			attachPendingRangeMarkers(table, pendingRangeMarkers);
-			content.push(table);
-			return;
-		}
-		if (localName === "AlternateContent") {
-			const selectedBranch = selectAlternateContentBranch(child);
-			if (!selectedBranch) return;
-			const branchOptions = withContainerXmlns(withContainerXmlns(childOptions, child), selectedBranch);
-			for (const selectedChild of getChildElements(selectedBranch)) parseCellChild(selectedChild, branchOptions);
-			return;
-		}
-		if (localName === "sdt") {
-			const sdtContent = findChildByLocalName(child, "sdtContent");
-			if (!sdtContent) return;
-			const sdtContentOptions = withContainerXmlns(withContainerXmlns(childOptions, child), sdtContent);
-			for (const sdtChild of getChildElements(sdtContent)) parseCellChild(sdtChild, sdtContentOptions);
-			return;
-		}
-		if (localName === "bookmarkStart" || localName === "bookmarkEnd") {
-			const marker = parseBookmarkMarker(child, localName);
-			if (!appendBookmarkMarkerToLastParagraphInBlocks(content, marker)) pendingBookmarkMarkers.push(marker);
-			return;
-		}
-		if (isBlockRangeMarker(localName)) pendingRangeMarkers.push(captureVerbatimXml(child));
+	const captured = [];
+	const dispatchCellChildren = (element, childOptions) => {
+		const preserved = dispatchChildren({
+			element,
+			container: "block-content",
+			modelledCount: () => modelled.length,
+			undeclared: { AlternateContent: (child) => {
+				const selectedBranch = selectAlternateContentBranch(child);
+				if (!selectedBranch) return;
+				dispatchCellChildren(selectedBranch, withContainerXmlns(withContainerXmlns(childOptions, child), selectedBranch));
+			} },
+			handlers: {
+				p: (child) => {
+					const para = parseParagraph(child, styles, theme, numbering, rels, media, childOptions);
+					enrichParagraphTextBoxes(para, child, styles, theme, numbering, rels, media, parseTable);
+					prependPendingBookmarkMarkers(para, pendingBookmarkMarkers);
+					modelled.push(para);
+				},
+				tbl: (child) => {
+					const table = parseTable(child, styles, theme, numbering, rels, media, childOptions);
+					if (!table) return;
+					if (prependBookmarkMarkersToFirstParagraphInBlocks([table], pendingBookmarkMarkers)) pendingBookmarkMarkers.length = 0;
+					modelled.push(table);
+				},
+				sdt: (child) => {
+					const sdtContent = findChildByLocalName(child, "sdtContent");
+					if (!sdtContent) return;
+					dispatchCellChildren(sdtContent, withContainerXmlns(withContainerXmlns(childOptions, child), sdtContent));
+				},
+				bookmarkStart: (child) => {
+					collectCellBookmarkMarker(child, "bookmarkStart", modelled, pendingBookmarkMarkers);
+				},
+				bookmarkEnd: (child) => {
+					collectCellBookmarkMarker(child, "bookmarkEnd", modelled, pendingBookmarkMarkers);
+				},
+				tcPr: OWNED_ELSEWHERE,
+				sectPr: CAPTURE,
+				altChunk: CAPTURE,
+				commentRangeEnd: CAPTURE,
+				commentRangeStart: CAPTURE,
+				customXml: CAPTURE,
+				customXmlDelRangeEnd: CAPTURE,
+				customXmlDelRangeStart: CAPTURE,
+				customXmlInsRangeEnd: CAPTURE,
+				customXmlInsRangeStart: CAPTURE,
+				customXmlMoveFromRangeEnd: CAPTURE,
+				customXmlMoveFromRangeStart: CAPTURE,
+				customXmlMoveToRangeEnd: CAPTURE,
+				customXmlMoveToRangeStart: CAPTURE,
+				del: CAPTURE,
+				ins: CAPTURE,
+				moveFrom: CAPTURE,
+				moveFromRangeEnd: CAPTURE,
+				moveFromRangeStart: CAPTURE,
+				moveTo: CAPTURE,
+				moveToRangeEnd: CAPTURE,
+				moveToRangeStart: CAPTURE,
+				permEnd: CAPTURE,
+				permStart: CAPTURE,
+				proofErr: CAPTURE
+			}
+		});
+		if (preserved?.children) captured.push(...preserved.children);
 	};
-	for (const child of elements) parseCellChild(child);
-	if (content.length === 0) content.push({
+	dispatchCellChildren(tcElement, options);
+	if (modelled.length === 0) modelled.push({
 		type: "paragraph",
 		content: [...pendingBookmarkMarkers]
 	});
-	else if (pendingBookmarkMarkers.length > 0) appendBookmarkMarkersToLastParagraphInBlocks(content, pendingBookmarkMarkers);
-	attachTrailingRangeMarkers(content, pendingRangeMarkers);
-	return content;
-}
+	else if (pendingBookmarkMarkers.length > 0) appendBookmarkMarkersToLastParagraphInBlocks(modelled, pendingBookmarkMarkers);
+	return withPreservedChildren(modelled, { children: captured }, (xml) => ({
+		type: "preservedBlock",
+		xml
+	}));
+}
+const collectCellBookmarkMarker = (child, localName, content, pending) => {
+	const marker = parseBookmarkMarker(child, localName);
+	if (!appendBookmarkMarkerToLastParagraphInBlocks(content, marker)) pending.push(marker);
+};
 /**
 * Parse a table cell (w:tc)
 *
@@ -576,29 +605,79 @@ function parseTableRow(trElement, styles, theme, numbering, rels, media, options
 	if (rowStructChange !== void 0) row.structuralChange = rowStructChange;
 	const rowOptions = withContainerXmlns(options, trElement);
 	const pendingBookmarkMarkers = [];
-	const parseRowChild = (child, childOptions = rowOptions) => {
-		const localName = getLocalName(child.name);
-		if (localName === "tc") {
-			const cell = parseTableCell(child, styles, theme, numbering, rels, media, childOptions);
-			if (pendingBookmarkMarkers.length > 0) {
-				prependBookmarkMarkersToFirstParagraphInCell(cell, pendingBookmarkMarkers);
-				pendingBookmarkMarkers.length = 0;
+	const preservedChildren = [];
+	/**
+	* One row's children, or a row-level content control's.
+	*
+	* folio unwraps `w:sdt` here and splices its rows' content into the row, so
+	* the recursion walks the control's content with the same map; the sink is
+	* the row's either way, because the control keeps no wrapper to hold one.
+	*/
+	const dispatchRowChildren = (element, childOptions) => {
+		const captured = dispatchChildren({
+			element,
+			container: "row-content",
+			modelledCount: () => row.cells.length,
+			handlers: {
+				tc: (child) => {
+					const cell = parseTableCell(child, styles, theme, numbering, rels, media, childOptions);
+					if (pendingBookmarkMarkers.length > 0) {
+						prependBookmarkMarkersToFirstParagraphInCell(cell, pendingBookmarkMarkers);
+						pendingBookmarkMarkers.length = 0;
+					}
+					row.cells.push(cell);
+				},
+				sdt: (child) => {
+					const sdtContent = findChildByLocalName(child, "sdtContent");
+					if (!sdtContent) return;
+					const sdtOptions = withContainerXmlns(childOptions, child);
+					dispatchRowChildren(sdtContent, withContainerXmlns(sdtOptions, sdtContent));
+				},
+				bookmarkStart: (child) => {
+					placeBookmarkMarker(parseBookmarkStart(child));
+				},
+				bookmarkEnd: (child) => {
+					placeBookmarkMarker(parseBookmarkEnd(child));
+				},
+				trPr: OWNED_ELSEWHERE,
+				tblPrEx: OWNED_ELSEWHERE,
+				commentRangeEnd: CAPTURE,
+				commentRangeStart: CAPTURE,
+				customXml: CAPTURE,
+				customXmlDelRangeEnd: CAPTURE,
+				customXmlDelRangeStart: CAPTURE,
+				customXmlInsRangeEnd: CAPTURE,
+				customXmlInsRangeStart: CAPTURE,
+				customXmlMoveFromRangeEnd: CAPTURE,
+				customXmlMoveFromRangeStart: CAPTURE,
+				customXmlMoveToRangeEnd: CAPTURE,
+				customXmlMoveToRangeStart: CAPTURE,
+				del: CAPTURE,
+				ins: CAPTURE,
+				moveFrom: CAPTURE,
+				moveFromRangeEnd: CAPTURE,
+				moveFromRangeStart: CAPTURE,
+				moveTo: CAPTURE,
+				moveToRangeEnd: CAPTURE,
+				moveToRangeStart: CAPTURE,
+				permEnd: CAPTURE,
+				permStart: CAPTURE,
+				proofErr: CAPTURE,
+				tr: CAPTURE
 			}
-			row.cells.push(cell);
-			return;
-		}
-		if (localName === "sdt") {
-			const sdtContent = findChildByLocalName(child, "sdtContent");
-			if (!sdtContent) return;
-			const sdtContentOptions = withContainerXmlns(withContainerXmlns(childOptions, child), sdtContent);
-			for (const sdtChild of getChildElements(sdtContent)) parseRowChild(sdtChild, sdtContentOptions);
-			return;
-		}
-		if (localName !== "bookmarkStart" && localName !== "bookmarkEnd") return;
-		const marker = parseBookmarkMarker(child, localName);
+		});
+		if (captured?.children) preservedChildren.push(...captured.children);
+	};
+	const placeBookmarkMarker = (marker) => {
 		if (!appendBookmarkMarkerToLastParagraphInCells(row.cells, marker)) pendingBookmarkMarkers.push(marker);
 	};
-	for (const child of getChildElements(trElement)) parseRowChild(child);
+	dispatchRowChildren(trElement, rowOptions);
+	if (preservedChildren.length > 0) row.preserved = { children: preservedChildren };
+	const remainingAttributes = attributeRemainder({
+		element: trElement,
+		modelled: NO_MODELLED_ATTRIBUTES
+	});
+	if (remainingAttributes) row.preservedAttributes = remainingAttributes;
 	if (pendingBookmarkMarkers.length > 0) appendBookmarkMarkersToLastCell(row, pendingBookmarkMarkers);
 	return row;
 }
@@ -707,23 +786,65 @@ function parseTable(tblElement, styles, theme, numbering, rels, media, options)
 	}
 	const tableOptions = withContainerXmlns(options, tblElement);
 	const rowsWithGridOffsets = /* @__PURE__ */ new Set();
-	const parseTableChild = (child, childOptions = tableOptions) => {
-		const localName = getLocalName(child.name);
-		if (localName === "tr") {
-			const rowIndex = table.rows.length;
-			const row = parseTableRow(child, styles, theme, numbering, rels, media, childOptions);
-			table.rows.push(row);
-			if (hasRowGridOffsets(child)) rowsWithGridOffsets.add(rowIndex);
-			return;
-		}
-		if (localName !== "sdt") return;
-		const sdtContent = findChildByLocalName(child, "sdtContent");
-		if (!sdtContent) return;
-		const sdtContentOptions = withContainerXmlns(withContainerXmlns(childOptions, child), sdtContent);
-		for (const sdtChild of getChildElements(sdtContent)) parseTableChild(sdtChild, sdtContentOptions);
+	const preservedChildren = [];
+	/**
+	* One table's children, or a row-level content control's.
+	*
+	* folio unwraps `w:sdt` here and splices its rows into the table, so the
+	* recursion walks the control's content with the same map; the sink is the
+	* table's either way, because the control keeps no wrapper to hold one.
+	*/
+	const dispatchTableChildren = (element, childOptions) => {
+		const captured = dispatchChildren({
+			element,
+			container: "table-content",
+			modelledCount: () => table.rows.length,
+			handlers: {
+				tr: (child) => {
+					const rowIndex = table.rows.length;
+					table.rows.push(parseTableRow(child, styles, theme, numbering, rels, media, childOptions));
+					if (hasRowGridOffsets(child)) rowsWithGridOffsets.add(rowIndex);
+				},
+				sdt: (child) => {
+					const sdtContent = findChildByLocalName(child, "sdtContent");
+					if (!sdtContent) return;
+					const sdtOptions = withContainerXmlns(childOptions, child);
+					dispatchTableChildren(sdtContent, withContainerXmlns(sdtOptions, sdtContent));
+				},
+				tblPr: OWNED_ELSEWHERE,
+				tblGrid: OWNED_ELSEWHERE,
+				bookmarkEnd: CAPTURE,
+				bookmarkStart: CAPTURE,
+				commentRangeEnd: CAPTURE,
+				commentRangeStart: CAPTURE,
+				customXml: CAPTURE,
+				customXmlPr: CAPTURE,
+				customXmlDelRangeEnd: CAPTURE,
+				customXmlDelRangeStart: CAPTURE,
+				customXmlInsRangeEnd: CAPTURE,
+				customXmlInsRangeStart: CAPTURE,
+				customXmlMoveFromRangeEnd: CAPTURE,
+				customXmlMoveFromRangeStart: CAPTURE,
+				customXmlMoveToRangeEnd: CAPTURE,
+				customXmlMoveToRangeStart: CAPTURE,
+				del: CAPTURE,
+				ins: CAPTURE,
+				moveFrom: CAPTURE,
+				moveFromRangeEnd: CAPTURE,
+				moveFromRangeStart: CAPTURE,
+				moveTo: CAPTURE,
+				moveToRangeEnd: CAPTURE,
+				moveToRangeStart: CAPTURE,
+				permEnd: CAPTURE,
+				permStart: CAPTURE,
+				proofErr: CAPTURE
+			}
+		});
+		if (captured?.children) preservedChildren.push(...captured.children);
 	};
-	for (const child of getChildElements(tblElement)) parseTableChild(child);
+	dispatchTableChildren(tblElement, tableOptions);
 	if (table.rows.length === 0) return;
+	if (preservedChildren.length > 0) table.preserved = { children: preservedChildren };
 	inferImplicitSingleCellRowSpans(table, rowsWithGridOffsets);
 	return table;
 }

package/dist/docx/textBoxParser.d.ts

@@ -1,7 +1,7 @@
 import { document_d_exports } from "../types/document.js";
 import { NumberingMap } from "./numberingParser.js";
-import { StyleMap } from "./styleParser.js";
 import { XmlElement } from "./xmlParser.js";
+import { StyleMap } from "./styleParser.js";
 //#region src/docx/textBoxParser.d.ts
 /**
  * Extract raw paragraph elements from w:txbxContent
@@ -112,5 +112,28 @@ declare function resolveTextBoxOutlineColor(textBox: document_d_exports.TextBox)
  * Get outline width in pixels
  */
 declare function getTextBoxOutlineWidthPx(textBox: document_d_exports.TextBox): number;
+type TextBoxRunScan = {
+  textBoxDrawings: XmlElement[];
+  vmlTextBoxes: XmlElement[];
+  hasNonTextBoxContent: boolean;
+};
+/**
+ * What a run's markup holds for the text-box pass, by kind.
+ *
+ * Two readers need this answer and they must not drift: the enrichment pass
+ * claims these elements, and `parseParagraphContents` has to keep the run they
+ * sit in even though the run parses to nothing until the pass fills it. It
+ * lives here, with `isTextBoxDrawing` and `parseTextBox`, so there is one.
+ */
+type ScanRunForTextBoxDrawingsOptions = {
+  xmlRun: XmlElement;
+  /**
+   * `isVmlPictParsedByRunParser`, passed rather than imported: the run parser
+   * reaches this module through the image parsers, so reading it from here
+   * would close a cycle.
+   */
+  claimedByRunParser: (pictElement: XmlElement) => boolean;
+};
+declare const scanRunForTextBoxDrawings: ({ xmlRun, claimedByRunParser }: ScanRunForTextBoxDrawingsOptions) => TextBoxRunScan;
 //#endregion
-export { ParagraphParserFn, TableParserFn, extractTextBoxContentElements, getTextBoxContentElement, getTextBoxDimensionsPx, getTextBoxHeightPx, getTextBoxMarginsPx, getTextBoxOutlineWidthPx, getTextBoxText, getTextBoxWidthPx, hasTextBoxContent, hasTextBoxFill, hasTextBoxOutline, isFloatingTextBox, isShapeTextBox, isTextBoxDrawing, parseTextBox, parseTextBoxContent, parseTextBoxFromShape, resolveTextBoxFillColor, resolveTextBoxOutlineColor };
+export { ParagraphParserFn, ScanRunForTextBoxDrawingsOptions, TableParserFn, TextBoxRunScan, extractTextBoxContentElements, getTextBoxContentElement, getTextBoxDimensionsPx, getTextBoxHeightPx, getTextBoxMarginsPx, getTextBoxOutlineWidthPx, getTextBoxText, getTextBoxWidthPx, hasTextBoxContent, hasTextBoxFill, hasTextBoxOutline, isFloatingTextBox, isShapeTextBox, isTextBoxDrawing, parseTextBox, parseTextBoxContent, parseTextBoxFromShape, resolveTextBoxFillColor, resolveTextBoxOutlineColor, scanRunForTextBoxDrawings };

package/dist/docx/textBoxParser.js

@@ -1,7 +1,7 @@
 import { emuToPixels } from "../utils/units.js";
 import { parseAnchorPosition, parseAnchorWrap, parseFill, parseOutline, resolveColorValueToHex } from "./drawingUtils.js";
 import { parseNonVisualDrawingNames } from "./nonVisualDrawingProps.js";
-import { findByFullName, findChildByLocalName, findChildByNamespaceUri, findChildrenByLocalName, findChildrenByNamespaceUri, getAttribute, getChildElements, parseNumericAttribute, parseOnOffAttribute } from "./xmlParser.js";
+import { findByFullName, findChildByLocalName, findChildByNamespaceUri, findChildrenByLocalName, findChildrenByNamespaceUri, findDeep, getAttribute, getChildElements, getLocalName, parseNumericAttribute, parseOnOffAttribute } from "./xmlParser.js";
 //#region src/docx/textBoxParser.ts
 const DRAWINGML_MAIN_NAMESPACE_URIS = /* @__PURE__ */ new Set(["http://schemas.openxmlformats.org/drawingml/2006/main", "http://purl.oclc.org/ooxml/drawingml/main"]);
 const WORDPROCESSING_SHAPE_NAMESPACE_URIS = /* @__PURE__ */ new Set(["http://schemas.microsoft.com/office/word/2010/wordprocessingShape"]);
@@ -290,6 +290,7 @@ const getTextBoxBlockText = (block) => {
 		return runTexts.join("");
 	}
 	if (block.type === "table") return block.rows.map((row) => row.cells.map((cell) => cell.content.map(getTextBoxBlockText).join("\n")).join("	")).join("\n");
+	if (block.type === "preservedBlock") return "";
 	return block.content.map(getTextBoxBlockText).join("\n");
 };
 /**
@@ -313,5 +314,54 @@ function getTextBoxOutlineWidthPx(textBox) {
 	if (!textBox.outline?.width) return 0;
 	return emuToPixels(textBox.outline.width);
 }
+const scanRunForTextBoxDrawings = ({ xmlRun, claimedByRunParser }) => {
+	const textBoxDrawings = [];
+	const vmlTextBoxes = [];
+	let hasNonTextBoxContent = false;
+	const visitDrawing = (drawingEl) => {
+		if (isTextBoxDrawing(drawingEl)) {
+			textBoxDrawings.push(drawingEl);
+			return;
+		}
+		hasNonTextBoxContent = true;
+	};
+	for (const el of getChildElements(xmlRun)) {
+		const name = getLocalName(el.name ?? "");
+		if (name === "rPr") continue;
+		if (name === "drawing") {
+			visitDrawing(el);
+			continue;
+		}
+		if (name === "pict") {
+			if (findDeep(el, "v", "textbox") && !claimedByRunParser(el)) vmlTextBoxes.push(el);
+			else hasNonTextBoxContent = true;
+			continue;
+		}
+		if (name === "AlternateContent") {
+			const branches = getChildElements(el);
+			const choice = branches.find((branch) => getLocalName(branch.name ?? "") === "Choice");
+			const fallback = branches.find((branch) => getLocalName(branch.name ?? "") === "Fallback");
+			const tryBranch = (branch) => {
+				if (!branch) return false;
+				let found = false;
+				for (const innerEl of getChildElements(branch)) if (getLocalName(innerEl.name ?? "") === "drawing") {
+					visitDrawing(innerEl);
+					found = true;
+				}
+				return found;
+			};
+			let foundInBranch = tryBranch(choice);
+			if (!foundInBranch) foundInBranch = tryBranch(fallback);
+			if (!foundInBranch) hasNonTextBoxContent = true;
+			continue;
+		}
+		hasNonTextBoxContent = true;
+	}
+	return {
+		textBoxDrawings,
+		vmlTextBoxes,
+		hasNonTextBoxContent
+	};
+};
 //#endregion
-export { extractTextBoxContentElements, getTextBoxContentElement, getTextBoxDimensionsPx, getTextBoxHeightPx, getTextBoxMarginsPx, getTextBoxOutlineWidthPx, getTextBoxText, getTextBoxWidthPx, hasTextBoxContent, hasTextBoxFill, hasTextBoxOutline, isFloatingTextBox, isShapeTextBox, isTextBoxDrawing, parseTextBox, parseTextBoxContent, parseTextBoxFromShape, resolveTextBoxFillColor, resolveTextBoxOutlineColor };
+export { extractTextBoxContentElements, getTextBoxContentElement, getTextBoxDimensionsPx, getTextBoxHeightPx, getTextBoxMarginsPx, getTextBoxOutlineWidthPx, getTextBoxText, getTextBoxWidthPx, hasTextBoxContent, hasTextBoxFill, hasTextBoxOutline, isFloatingTextBox, isShapeTextBox, isTextBoxDrawing, parseTextBox, parseTextBoxContent, parseTextBoxFromShape, resolveTextBoxFillColor, resolveTextBoxOutlineColor, scanRunForTextBoxDrawings };

package/dist/docx/unzip.js

@@ -1,6 +1,7 @@
 import { bytesToDataUrl } from "../utils/base64.js";
 import { DOCX_CONTAINER_TYPES, detectDocxContainerType } from "./encryption/containerFormat.js";
 import { openDocxBuffer } from "./encryption/openEncryptedDocx.js";
+import { decodeXmlBytes } from "./xmlEncoding.js";
 import { FOLIO_XML_RESOURCE_LIMITS, assertXmlResourceLimits, createXmlPackageBudget } from "./xmlResourceLimits.js";
 import JSZip from "jszip";
 //#region src/docx/unzip.ts
@@ -161,13 +162,13 @@ async function unzipDocx(buffer, options = {}) {
 		if (lowerPath.endsWith(".xml") || lowerPath.endsWith(".rels")) {
 			assertEntrySize(path, declaredSize, limits.maxXmlBytes);
 			if (options.extractAllXml === false && !shouldExtractXmlPart(lowerPath)) continue;
-			extractionTasks.push(() => file.async("text").then((xmlContent) => {
-				assertExtractedSize(path, xmlContent.length, limits.maxXmlBytes);
+			extractionTasks.push(() => file.async("uint8array").then((xmlBytes) => {
+				assertExtractedSize(path, xmlBytes.byteLength, limits.maxXmlBytes);
 				return {
 					type: "xml",
 					path,
 					lowerPath,
-					content: xmlContent
+					content: decodeXmlBytes(xmlBytes)
 				};
 			}));
 		} else if (lowerPath.startsWith("word/media/")) {
@@ -456,7 +457,7 @@ function extractFile(content, path) {
 	const file = content.originalZip.file(path);
 	if (!file || !isParsedDocxEntry(path)) return Promise.resolve(null);
 	const lowerPath = path.toLowerCase();
-	if (lowerPath.endsWith(".xml") || lowerPath.endsWith(".rels")) return file.async("text");
+	if (lowerPath.endsWith(".xml") || lowerPath.endsWith(".rels")) return file.async("uint8array").then(decodeXmlBytes);
 	return file.async("arraybuffer");
 }
 /**

package/dist/docx/vmlImageParser.d.ts

@@ -11,6 +11,22 @@ import { XmlElement } from "./xmlParser.js";
  * same artwork twice.
  */
 declare function shouldPreserveRawVmlPict(pictElement: XmlElement): boolean;
+/**
+ * Whether the run parser turns this `w:pict` into run content of its own.
+ *
+ * Exactly one owner may represent a `w:pict`: the run parser, whose drawing
+ * carries the whole element as captured XML, or the text-box enrichment pass,
+ * which rebuilds one `v:textbox` as an editable shape. Two owners write the
+ * same artwork twice, and when the pict holds text, the saved document says it
+ * twice.
+ *
+ * The enrichment asks this rather than re-deriving the answer from the markup:
+ * its own reading (`v:imagedata` present) named only the picture path, so a
+ * `v:group` claimed by the preview path was represented by both owners. The
+ * result is a pure function of the element, so asking it a second time costs
+ * only the work the run parser already did.
+ */
+declare const isVmlPictParsedByRunParser: (pictElement: XmlElement, rels: document_d_exports.RelationshipMap | null, media: Map<string, document_d_exports.MediaFile> | null) => boolean;
 /**
  * Parse a `w:pict` element into an inline image, or null when it carries no
  * ordinary VML picture (no resolvable `<v:imagedata>`, or a watermark shape).
@@ -22,4 +38,4 @@ declare function shouldPreserveRawVmlPict(pictElement: XmlElement): boolean;
  */
 declare function parseVmlImageContent(pictElement: XmlElement, rels: document_d_exports.RelationshipMap | null, media: Map<string, document_d_exports.MediaFile> | null, rootXmlns?: Record<string, string>): document_d_exports.DrawingContent | null;
 //#endregion
-export { parseVmlImageContent, shouldPreserveRawVmlPict };
+export { isVmlPictParsedByRunParser, parseVmlImageContent, shouldPreserveRawVmlPict };

package/dist/docx/vmlImageParser.js

@@ -131,6 +131,22 @@ function shouldPreserveRawVmlPict(pictElement) {
 	return shapes.length > 0 && !shapes.some((shape) => isWatermarkShape(shape));
 }
 /**
+* Whether the run parser turns this `w:pict` into run content of its own.
+*
+* Exactly one owner may represent a `w:pict`: the run parser, whose drawing
+* carries the whole element as captured XML, or the text-box enrichment pass,
+* which rebuilds one `v:textbox` as an editable shape. Two owners write the
+* same artwork twice, and when the pict holds text, the saved document says it
+* twice.
+*
+* The enrichment asks this rather than re-deriving the answer from the markup:
+* its own reading (`v:imagedata` present) named only the picture path, so a
+* `v:group` claimed by the preview path was represented by both owners. The
+* result is a pure function of the element, so asking it a second time costs
+* only the work the run parser already did.
+*/
+const isVmlPictParsedByRunParser = (pictElement, rels, media) => parseVmlImageContent(pictElement, rels, media) !== null || shouldPreserveRawVmlPict(pictElement);
+/**
 * Read the relationship id off a `v:imagedata` element. Word writes `r:id`;
 * some legacy / third-party generators use `r:embed` or the office-namespace
 * `o:relid` instead, so fall back through those before the bare `id`.
@@ -201,4 +217,4 @@ function parseVmlImageContent(pictElement, rels, media, rootXmlns = {}) {
 	return null;
 }
 //#endregion
-export { parseVmlImageContent, shouldPreserveRawVmlPict };
+export { isVmlPictParsedByRunParser, parseVmlImageContent, shouldPreserveRawVmlPict };

package/dist/docx/xmlEncoding.d.ts

@@ -0,0 +1,5 @@
+//#region src/docx/xmlEncoding.d.ts
+/** Decode the UTF-8 and UTF-16 encodings XML processors must recognize. */
+declare const decodeXmlBytes: (bytes: Uint8Array) => string;
+//#endregion
+export { decodeXmlBytes };

package/dist/docx/xmlEncoding.js

@@ -0,0 +1,12 @@
+//#region src/docx/xmlEncoding.ts
+/** Decode the UTF-8 and UTF-16 encodings XML processors must recognize. */
+const decodeXmlBytes = (bytes) => {
+	if (bytes[0] === 255 && bytes[1] === 254) return new TextDecoder("utf-16le").decode(bytes.subarray(2));
+	if (bytes[0] === 254 && bytes[1] === 255) return new TextDecoder("utf-16be").decode(bytes.subarray(2));
+	if (bytes[0] === 60 && bytes[1] === 0 && bytes[2] === 63 && bytes[3] === 0) return new TextDecoder("utf-16le").decode(bytes);
+	if (bytes[0] === 0 && bytes[1] === 60 && bytes[2] === 0 && bytes[3] === 63) return new TextDecoder("utf-16be").decode(bytes);
+	const utf8Offset = bytes[0] === 239 && bytes[1] === 187 && bytes[2] === 191 ? 3 : 0;
+	return new TextDecoder("utf-8").decode(bytes.subarray(utf8Offset));
+};
+//#endregion
+export { decodeXmlBytes };

package/dist/docx/xmlParser.d.ts

@@ -105,6 +105,15 @@ declare function getLocalName(name: string | undefined): string;
 declare function getNamespacePrefix(name: string): string | null;
 /** Namespace URI resolved from the element's in-scope XML declarations. */
 declare const getNamespaceUri: (element: XmlElement) => string | undefined;
+/**
+ * Namespace URI of one of the element's attributes, by its source spelling.
+ *
+ * An unprefixed attribute has no namespace — it is not in the element's,
+ * which is why this is not {@link getNamespaceUri} with a different argument —
+ * and `undefined` is that answer as well as "the prefix resolves to nothing".
+ * A caller that must tell the two apart checks the spelling for a colon.
+ */
+declare const resolveAttributeNamespaceUri: (element: XmlElement, attributeName: string) => string | undefined;
 /** WordprocessingML main namespace, Transitional and Strict (ECMA-376 Parts 1 and 4). */
 declare const WORDPROCESSINGML_NAMESPACE_URIS: ReadonlySet<string>;
 /** Office document relationship attributes, Transitional and Strict. */
@@ -383,4 +392,4 @@ declare function cloneWithXmlnsDeclarations(element: XmlElement, xmlnsDecls: Rec
  */
 declare function cloneElement(element: XmlElement, overrides: Partial<XmlElement>): XmlElement;
 //#endregion
-export { NAMESPACES, OFFICE_RELATIONSHIP_NAMESPACE_URIS, OOXML_NAMESPACE_SCOPE, WORDPROCESSINGML_NAMESPACE_URIS, XmlAttributeMatch, XmlElement, XmlNamespaceScope, attachXmlNamespaceContext, cloneElement, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findChildrenByNamespaceUri, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument, selectAlternateContentBranch };
+export { NAMESPACES, OFFICE_RELATIONSHIP_NAMESPACE_URIS, OOXML_NAMESPACE_SCOPE, WORDPROCESSINGML_NAMESPACE_URIS, XmlAttributeMatch, XmlElement, XmlNamespaceScope, attachXmlNamespaceContext, cloneElement, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findChildrenByNamespaceUri, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument, resolveAttributeNamespaceUri, selectAlternateContentBranch };

package/dist/docx/xmlParser.js

@@ -246,6 +246,19 @@ function getNamespacePrefix(name) {
 }
 /** Namespace URI resolved from the element's in-scope XML declarations. */
 const getNamespaceUri = (element) => element.namespaceUri;
+/**
+* Namespace URI of one of the element's attributes, by its source spelling.
+*
+* An unprefixed attribute has no namespace — it is not in the element's,
+* which is why this is not {@link getNamespaceUri} with a different argument —
+* and `undefined` is that answer as well as "the prefix resolves to nothing".
+* A caller that must tell the two apart checks the spelling for a colon.
+*/
+const resolveAttributeNamespaceUri = (element, attributeName) => {
+	const colonIndex = attributeName.indexOf(":");
+	if (colonIndex === -1) return;
+	return resolveNamespaceUri(element.namespaceScope, attributeName.slice(0, colonIndex));
+};
 /** WordprocessingML main namespace, Transitional and Strict (ECMA-376 Parts 1 and 4). */
 const WORDPROCESSINGML_NAMESPACE_URIS = /* @__PURE__ */ new Set([NAMESPACES.w, "http://purl.oclc.org/ooxml/wordprocessingml/main"]);
 /** Office document relationship attributes, Transitional and Strict. */
@@ -855,4 +868,4 @@ function cloneElement(element, overrides) {
 	return clone;
 }
 //#endregion
-export { NAMESPACES, OFFICE_RELATIONSHIP_NAMESPACE_URIS, OOXML_NAMESPACE_SCOPE, WORDPROCESSINGML_NAMESPACE_URIS, attachXmlNamespaceContext, cloneElement, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findChildrenByNamespaceUri, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument, selectAlternateContentBranch };
+export { NAMESPACES, OFFICE_RELATIONSHIP_NAMESPACE_URIS, OOXML_NAMESPACE_SCOPE, WORDPROCESSINGML_NAMESPACE_URIS, attachXmlNamespaceContext, cloneElement, cloneWithXmlnsDeclarations, collectXmlnsDeclarations, elementToXml, findAllDeep, findAttributeByNamespaceUri, findByFullName, findChild, findChildByLocalName, findChildByNamespaceUri, findChildren, findChildrenByLocalName, findChildrenByNamespaceUri, findDeep, getAttribute, getAttributeAny, getAttributeAnyPrefix, getAttributeByNamespaceUri, getAttributes, getChildElements, getLocalName, getNamespacePrefix, getNamespaceUri, getTextContent, hasChild, matchesName, mergeXmlnsDeclarations, parseBooleanElement, parseColorElement, parseNumberingLevelAttribute, parseNumericAttribute, parseOnOffAttribute, parseOnOffValue, parseTableMeasurementValue, parseXml, parseXmlDocument, resolveAttributeNamespaceUri, selectAlternateContentBranch };

package/dist/headless-layout.js

@@ -1,5 +1,7 @@
 import { resolveDocumentGridLinePitch } from "./docx/documentGrid.js";
 import { formatOoxmlCounter } from "./docx/ooxmlCounterFormatter.js";
+import { createParseWarningCollector } from "./docx/parseContext.js";
+import { formatParseWarnings } from "./docx/parseWarningMessage.js";
 import { parseDocx } from "./docx/parser.js";
 import { buildHeaderFooterFieldValues } from "./fields/resolveFieldValues.js";
 import { extractEmbeddedFonts } from "./fonts/embeddedFonts.js";
@@ -209,14 +211,21 @@ const layoutDocxHeadless = async (input, options = {}) => {
 	});
 	if (parsed.isErr()) return Result.err(parsed.error);
 	const document = parsed.value;
+	const { context: projectionContext, warnings: projectionWarnings } = createParseWarningCollector("word/document.xml");
 	const projected = Result.try(() => toProseDoc(document, {
 		...document.package.styles ? { styles: document.package.styles } : {},
-		...document.package.theme ? { theme: document.package.theme } : {}
+		...document.package.theme ? { theme: document.package.theme } : {},
+		warn: projectionContext.warn
 	}));
 	if (projected.isErr()) return Result.err(new HeadlessLayoutError({
 		message: "The document could not be projected to the editable model.",
 		cause: projected.error
 	}));
+	const recorded = projectionWarnings();
+	if (recorded.length > 0) {
+		document.parseWarnings = [...document.parseWarnings ?? [], ...recorded];
+		document.warnings = [...document.warnings ?? [], ...formatParseWarnings(recorded)];
+	}
 	if (packageNeedsShaping(document.package)) {
 		const loaded = await Result.tryPromise({
 			try: () => getShaper(),