@stll/folio-core 0.43.0 → 0.45.0

package/dist/docx/defaultParagraphStyle.d.ts

@@ -17,6 +17,23 @@ declare const BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING: {
   lineSpacing: number;
   lineSpacingRule: "auto";
 };
+type MintDefaultParagraphStyleOptions = {
+  takenStyleIds: ReadonlySet<string>;
+  /** Whether the source declared `w:docDefaults`, which the set carries over. */
+  hasDocDefaults: boolean;
+};
+/**
+ * The default paragraph style a set needs when its source declared none.
+ *
+ * The id only has to be free, because the set is what defines it. The
+ * formatting has to be the built-in template's whenever the source had no
+ * `w:docDefaults`, because that is what the source itself rendered as: a
+ * consumer applies its built-in Normal only where no default paragraph style
+ * exists, and this minted style is one. Where the source did declare
+ * `w:docDefaults`, the set carries them and they remain authoritative, so the
+ * minted style states nothing.
+ */
+declare const mintDefaultParagraphStyle: ({ takenStyleIds, hasDocDefaults }: MintDefaultParagraphStyleOptions) => document_d_exports.Style;
 declare const resolveDefaultParagraphStyle: (styles: Iterable<document_d_exports.Style>) => document_d_exports.Style | undefined;
 //#endregion
-export { BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_ID, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_NAME, resolveDefaultParagraphStyle };
+export { BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_ID, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_NAME, mintDefaultParagraphStyle, resolveDefaultParagraphStyle };

package/dist/docx/defaultParagraphStyle.js

@@ -16,6 +16,28 @@ const BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING = {
 	lineSpacing: 259,
 	lineSpacingRule: "auto"
 };
+/**
+* The default paragraph style a set needs when its source declared none.
+*
+* The id only has to be free, because the set is what defines it. The
+* formatting has to be the built-in template's whenever the source had no
+* `w:docDefaults`, because that is what the source itself rendered as: a
+* consumer applies its built-in Normal only where no default paragraph style
+* exists, and this minted style is one. Where the source did declare
+* `w:docDefaults`, the set carries them and they remain authoritative, so the
+* minted style states nothing.
+*/
+const mintDefaultParagraphStyle = ({ takenStyleIds, hasDocDefaults }) => {
+	let styleId = BUILT_IN_DEFAULT_PARAGRAPH_STYLE_ID;
+	for (let suffix = 1; takenStyleIds.has(styleId); suffix += 1) styleId = `${BUILT_IN_DEFAULT_PARAGRAPH_STYLE_ID}${suffix}`;
+	return {
+		styleId,
+		type: "paragraph",
+		name: BUILT_IN_DEFAULT_PARAGRAPH_STYLE_NAME,
+		default: true,
+		...hasDocDefaults ? {} : { pPr: { ...BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING } }
+	};
+};
 const resolveDefaultParagraphStyle = (styles) => {
 	let flagged;
 	let namedBuiltIn;
@@ -29,4 +51,4 @@ const resolveDefaultParagraphStyle = (styles) => {
 	return flagged ?? namedBuiltIn ?? idBuiltIn;
 };
 //#endregion
-export { BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_ID, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_NAME, resolveDefaultParagraphStyle };
+export { BUILT_IN_DEFAULT_PARAGRAPH_FORMATTING, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_ID, BUILT_IN_DEFAULT_PARAGRAPH_STYLE_NAME, mintDefaultParagraphStyle, resolveDefaultParagraphStyle };

package/dist/docx/diagramPreview.js

@@ -1,5 +1,7 @@
-import { resolveRelativePath } from "./relsParser.js";
-import { findChildByNamespaceUri, getAttribute, getLocalName, parseNumericAttribute, parseXmlDocument } from "./xmlParser.js";
+import { bytesToDataUrl } from "../utils/base64.js";
+import { PREVIEW_KINDS } from "./previewBudget.js";
+import { resolveRelationshipIdOfType, resolveRelativePath } from "./relsParser.js";
+import { OFFICE_RELATIONSHIP_NAMESPACE_URIS, findChildByNamespaceUri, getAttribute, getAttributeByNamespaceUri, getLocalName, parseNumericAttribute, parseXmlDocument } from "./xmlParser.js";
 //#region src/docx/diagramPreview.ts
 const MAX_PREVIEW_SHAPES = 128;
 const MAX_PREVIEW_PIXELS = 144e4;
@@ -7,21 +9,47 @@ const MAX_PREVIEW_PAINT_PIXELS = MAX_PREVIEW_PIXELS * 4;
 const DRAWINGML_NAMESPACE_URIS = /* @__PURE__ */ new Set(["http://schemas.openxmlformats.org/drawingml/2006/main", "http://purl.oclc.org/ooxml/drawingml/main"]);
 const DIAGRAM_NAMESPACE_URIS = /* @__PURE__ */ new Set(["http://schemas.openxmlformats.org/drawingml/2006/diagram", "http://purl.oclc.org/ooxml/drawingml/diagram"]);
 const DIAGRAM_DRAWING_NAMESPACE_URIS = /* @__PURE__ */ new Set(["http://schemas.microsoft.com/office/drawing/2008/diagram"]);
+const DIAGRAM_DATA_RELATIONSHIP_TYPE = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/diagramData";
+const DIAGRAM_DRAWING_RELATIONSHIP_TYPE = "http://schemas.microsoft.com/office/2007/relationships/diagramDrawing";
+const DOCUMENT_PART_PATH = "word/document.xml";
 const WORD_DRAWING_NAMESPACE_URIS = /* @__PURE__ */ new Set(["http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing", "http://purl.oclc.org/ooxml/drawingml/wordprocessingDrawing"]);
+/**
+* The preview is a megapixel raster, so both checksums run over megabytes.
+* `for (const byte of bytes)` drives the array iterator protocol once per
+* byte, which profiles as the dominant cost of parsing a SmartArt document;
+* indexed loops and a table-driven CRC produce the same numbers without it.
+*/
+const CRC32_TABLE = (() => {
+	const table = /* @__PURE__ */ new Uint32Array(256);
+	for (let index = 0; index < 256; index += 1) {
+		let value = index;
+		for (let bit = 0; bit < 8; bit += 1) value = value >>> 1 ^ (value & 1 ? 3988292384 : 0);
+		table[index] = value >>> 0;
+	}
+	return table;
+})();
 const crc32 = (bytes) => {
 	let crc = 4294967295;
-	for (const byte of bytes) {
-		crc ^= byte;
-		for (let bit = 0; bit < 8; bit += 1) crc = crc >>> 1 ^ (crc & 1 ? 3988292384 : 0);
-	}
+	for (let index = 0; index < bytes.length; index += 1) crc = crc >>> 8 ^ CRC32_TABLE[(crc ^ bytes[index]) & 255];
 	return (crc ^ 4294967295) >>> 0;
 };
+/**
+* `a` and `b` stay below 2^31 for 5552 iterations from any legal state, so the
+* modulo runs per block rather than per byte.
+*/
+const ADLER32_BLOCK = 5552;
 const adler32 = (bytes) => {
 	let a = 1;
 	let b = 0;
-	for (const byte of bytes) {
-		a = (a + byte) % 65521;
-		b = (b + a) % 65521;
+	let index = 0;
+	while (index < bytes.length) {
+		const end = Math.min(index + ADLER32_BLOCK, bytes.length);
+		for (; index < end; index += 1) {
+			a += bytes[index];
+			b += a;
+		}
+		a %= 65521;
+		b %= 65521;
 	}
 	return b << 16 | a;
 };
@@ -88,9 +116,8 @@ const previewPng = (width, height, shapes) => {
 		cursor += length;
 		offset += length;
 	}
-	const output = new Uint8Array(cursor + 4);
-	output.set(compressed.subarray(0, cursor));
-	new DataView(output.buffer).setUint32(cursor, adler32(pixels));
+	new DataView(compressed.buffer).setUint32(cursor, adler32(pixels));
+	const output = compressed;
 	const signature = new Uint8Array([
 		137,
 		80,
@@ -127,13 +154,38 @@ const extent = (drawing) => {
 		height: parseNumericAttribute(value, null, "cy") ?? 0
 	};
 };
-const cachedDiagramShapes = (rels, media) => {
-	const matches = [...rels.values()].filter((relationship) => relationship.type === "http://schemas.microsoft.com/office/2007/relationships/diagramDrawing");
-	if (matches.length !== 1 || !matches[0]?.target) return [];
-	const path = resolveRelativePath("word/document.xml", matches[0].target);
-	const file = media.get(path);
-	if (!file?.data) return [];
-	const root = parseXmlDocument(new TextDecoder().decode(file.data));
+/** Parse the part a relationship of the given type names, by that relationship's id. */
+const partByRelationshipId = ({ rels, media, rId, type }) => {
+	const resolved = resolveRelationshipIdOfType(rels, rId ?? void 0, type);
+	if (resolved.status !== "resolved" || !resolved.relationship.target) return null;
+	const file = media.get(resolveRelativePath(DOCUMENT_PART_PATH, resolved.relationship.target));
+	return file?.data ? parseXmlDocument(new TextDecoder().decode(file.data)) : null;
+};
+/**
+* The drawing cache this diagram points at, reached through its own ids.
+*
+* `dgm:relIds/@r:dm` names the data part, and that part's `dsp:dataModelExt`
+* extension names the cached drawing. Scanning the relationship map for a type
+* instead of following the ids reads the wrong diagram whenever a document has
+* more than one, which is why the scan refused outright on a second match:
+* both diagrams in a two-diagram document then got no preview at all.
+*/
+const cachedDiagramShapes = (graphicData, rels, media) => {
+	const relIds = findChildByNamespaceUri(graphicData, DIAGRAM_NAMESPACE_URIS, "relIds");
+	const data = partByRelationshipId({
+		rels,
+		media,
+		rId: getAttributeByNamespaceUri(relIds, OFFICE_RELATIONSHIP_NAMESPACE_URIS, "dm"),
+		type: DIAGRAM_DATA_RELATIONSHIP_TYPE
+	});
+	if (!data) return [];
+	const dataModelExt = descendantsByNamespace(data, DIAGRAM_DRAWING_NAMESPACE_URIS, "dataModelExt").at(0);
+	const root = partByRelationshipId({
+		rels,
+		media,
+		rId: getAttribute(dataModelExt, null, "relId"),
+		type: DIAGRAM_DRAWING_RELATIONSHIP_TYPE
+	});
 	if (!root) return [];
 	const shapes = [];
 	for (const shape of descendantsByNamespace(root, DIAGRAM_DRAWING_NAMESPACE_URIS, "sp").slice(0, MAX_PREVIEW_SHAPES)) {
@@ -155,31 +207,39 @@ const cachedDiagramShapes = (rels, media) => {
 	}
 	return shapes;
 };
+const matchesNamespace = (element, namespaceUris, localName) => element.namespaceUri !== void 0 && namespaceUris.has(element.namespaceUri) && getLocalName(element.name ?? "") === localName;
 const descendantsByNamespace = (root, namespaceUris, localName) => {
 	const result = [];
 	const visit = (element) => {
-		if (element.namespaceUri && namespaceUris.has(element.namespaceUri) && getLocalName(element.name ?? "") === localName) result.push(element);
+		if (matchesNamespace(element, namespaceUris, localName)) result.push(element);
 		for (const child of element.elements ?? []) if (child.type === "element") visit(child);
 	};
 	visit(root);
 	return result;
 };
+/** The first match in document order, without walking the rest of the subtree. */
+const firstDescendantByNamespace = (root, namespaceUris, localName) => {
+	if (matchesNamespace(root, namespaceUris, localName)) return root;
+	for (const child of root.elements ?? []) {
+		if (child.type !== "element") continue;
+		const found = firstDescendantByNamespace(child, namespaceUris, localName);
+		if (found) return found;
+	}
+	return null;
+};
 /** Create a deliberately simple, bounded preview; it is never an editable diagram projection. */
 const parseDiagramPreview = (drawing, rels, media) => {
 	if (!rels || !media) return null;
-	const graphicData = descendantsByNamespace(drawing, DRAWINGML_NAMESPACE_URIS, "graphicData").at(0);
+	const graphicData = firstDescendantByNamespace(drawing, DRAWINGML_NAMESPACE_URIS, "graphicData");
 	if (!graphicData || !DIAGRAM_NAMESPACE_URIS.has(getAttribute(graphicData, null, "uri") ?? "")) return null;
 	const { width, height } = extent(drawing);
 	if (width <= 0 || height <= 0) return null;
-	const png = previewPng(width, height, cachedDiagramShapes(rels, media));
-	let binary = "";
-	for (let offset = 0; offset < png.length; offset += 32768) binary += String.fromCodePoint(...png.subarray(offset, offset + 32768));
 	return {
 		type: "image",
 		rId: "",
-		src: `data:image/png;base64,${btoa(binary)}`,
-		mimeType: "image/png",
-		filename: "smartart-preview.png",
+		src: bytesToDataUrl(previewPng(width, height, cachedDiagramShapes(graphicData, rels, media)), PREVIEW_KINDS.smartArt.mimeType),
+		mimeType: PREVIEW_KINDS.smartArt.mimeType,
+		filename: PREVIEW_KINDS.smartArt.filename,
 		size: {
 			width,
 			height

package/dist/docx/documentParser.d.ts

@@ -1,6 +1,7 @@
 import { document_d_exports } from "../types/document.js";
 import { NumberingMap } from "./numberingParser.js";
 import { StyleMap } from "./styleParser.js";
+import { ParseContext } from "./parseContext.js";
 //#region src/docx/documentParser.d.ts
 /**
  * Extract template variables from text
@@ -27,7 +28,7 @@ declare function extractAllTemplateVariables(content: document_d_exports.BlockCo
  * @param media - Media files
  * @returns DocumentBody with content, sections, and template variables
  */
-declare function parseDocumentBody(xml: string, styles?: StyleMap | null, theme?: document_d_exports.Theme | null, numbering?: NumberingMap | null, rels?: document_d_exports.RelationshipMap | null, media?: Map<string, document_d_exports.MediaFile> | null): document_d_exports.DocumentBody;
+declare function parseDocumentBody(xml: string, styles?: StyleMap | null, theme?: document_d_exports.Theme | null, numbering?: NumberingMap | null, rels?: document_d_exports.RelationshipMap | null, media?: Map<string, document_d_exports.MediaFile> | null, context?: ParseContext): document_d_exports.DocumentBody;
 /**
  * Get all paragraphs from document body (flattened)
  */

package/dist/docx/documentParser.js

@@ -119,7 +119,7 @@ function canonicalizeLeadingBodySectionProperties(bodyElement, body) {
 * @param media - Media files
 * @returns DocumentBody with content, sections, and template variables
 */
-function parseDocumentBody(xml, styles = null, theme = null, numbering = null, rels = null, media = null) {
+function parseDocumentBody(xml, styles = null, theme = null, numbering = null, rels = null, media = null, context) {
 	const result = { content: [] };
 	if (!xml) return result;
 	const streamed = parseStreamingXml(xml);
@@ -129,7 +129,7 @@ function parseDocumentBody(xml, styles = null, theme = null, numbering = null, r
 	if (!bodyEl) return result;
 	result.content = parseBlockContent(bodyEl, styles, theme, numbering, rels, media, { rootXmlns: collectXmlnsDeclarations(documentEl) });
 	const finalSectPr = findChild(bodyEl, "w", "sectPr");
-	if (finalSectPr) result.finalSectionProperties = parseSectionProperties(finalSectPr);
+	if (finalSectPr) result.finalSectionProperties = parseSectionProperties(finalSectPr, context);
 	canonicalizeLeadingBodySectionProperties(bodyEl, result);
 	result.sections = buildSections(result.content, result.finalSectionProperties);
 	return result;

package/dist/docx/drawingUtils.d.ts

@@ -67,6 +67,13 @@ declare function parseWrapElement(wrapEl: XmlElement | null, behindDoc: boolean,
 /**
  * Parse wrap from an anchor element (finds wrap child internally).
  */
+/**
+ * Read `wp:anchor/@behindDoc`, the flag that puts an anchored object behind the
+ * body text. The attribute is xsd:boolean, so `1`, `0`, `true` and `false` are
+ * all legal spellings and producers differ: Word writes `1`, others `true`.
+ * Absent means in front of the text.
+ */
+declare function parseAnchorBehindDoc(anchor: XmlElement): boolean;
 declare function parseAnchorWrap(anchor: XmlElement): document_d_exports.ImageWrap | undefined;
 /**
  * Resolve a ColorValue to a CSS hex string using default theme colors.
@@ -74,4 +81,4 @@ declare function parseAnchorWrap(anchor: XmlElement): document_d_exports.ImageWr
  */
 declare function resolveColorValueToHex(color: document_d_exports.ColorValue | undefined): string | undefined;
 //#endregion
-export { THEME_COLOR_TO_DRAWING_SCHEME, WRAP_ELEMENT_NAMES, parseAnchorPosition, parseAnchorWrap, parseColorElement, parseFill, parseOutline, parsePositionH, parsePositionV, parseWrapElement, resolveColorValueToHex };
+export { THEME_COLOR_TO_DRAWING_SCHEME, WRAP_ELEMENT_NAMES, parseAnchorBehindDoc, parseAnchorPosition, parseAnchorWrap, parseColorElement, parseFill, parseOutline, parsePositionH, parsePositionV, parseWrapElement, resolveColorValueToHex };

package/dist/docx/drawingUtils.js

@@ -1,6 +1,6 @@
 import { ImageHorizontalAlignmentSchema, ImageHorizontalRelativeToSchema, ImageVerticalAlignmentSchema, ImageVerticalRelativeToSchema, ImageWrapTextSchema, ShapeOutlineStyleSchema, narrowEnum } from "./parserEnums.js";
 import { captureVerbatimXml } from "./verbatimCapture.js";
-import { findByFullName, findChildByLocalName, findChildrenByLocalName, getAttribute, getChildElements, getTextContent, parseNumericAttribute } from "./xmlParser.js";
+import { findByFullName, findChildByLocalName, findChildrenByLocalName, getAttribute, getChildElements, getTextContent, parseNumericAttribute, parseOnOffValue } from "./xmlParser.js";
 //#region src/docx/drawingUtils.ts
 /**
 * Map OOXML scheme names to standard theme color slots.
@@ -361,9 +361,18 @@ function parseWrapElement(wrapEl, behindDoc, anchorDistances) {
 /**
 * Parse wrap from an anchor element (finds wrap child internally).
 */
+/**
+* Read `wp:anchor/@behindDoc`, the flag that puts an anchored object behind the
+* body text. The attribute is xsd:boolean, so `1`, `0`, `true` and `false` are
+* all legal spellings and producers differ: Word writes `1`, others `true`.
+* Absent means in front of the text.
+*/
+function parseAnchorBehindDoc(anchor) {
+	return parseOnOffValue(getAttribute(anchor, null, "behindDoc")) ?? false;
+}
 function parseAnchorWrap(anchor) {
 	const children = getChildElements(anchor);
-	const behindDoc = getAttribute(anchor, null, "behindDoc") === "1";
+	const behindDoc = parseAnchorBehindDoc(anchor);
 	const wrapEl = children.find((el) => WRAP_ELEMENT_NAMES.includes(el.name ?? ""));
 	const distT = parseNumericAttribute(anchor, null, "distT");
 	const distB = parseNumericAttribute(anchor, null, "distB");
@@ -409,4 +418,4 @@ function resolveColorValueToHex(color) {
 	if (color.themeColor) return `#${DEFAULT_THEME_COLOR_HEX[color.themeColor] ?? "000000"}`;
 }
 //#endregion
-export { THEME_COLOR_TO_DRAWING_SCHEME, WRAP_ELEMENT_NAMES, parseAnchorPosition, parseAnchorWrap, parseColorElement, parseFill, parseOutline, parsePositionH, parsePositionV, parseWrapElement, resolveColorValueToHex };
+export { THEME_COLOR_TO_DRAWING_SCHEME, WRAP_ELEMENT_NAMES, parseAnchorBehindDoc, parseAnchorPosition, parseAnchorWrap, parseColorElement, parseFill, parseOutline, parsePositionH, parsePositionV, parseWrapElement, resolveColorValueToHex };

package/dist/docx/fieldParser.js

@@ -1,7 +1,7 @@
 import { formatOoxmlCounter } from "./ooxmlCounterFormatter.js";
 import { FieldTypeSchema, narrowEnum } from "./parserEnums.js";
 import { parseRun } from "./runParser.js";
-import { findChildren, getAttribute } from "./xmlParser.js";
+import { findChildren, getAttribute, parseOnOffAttribute } from "./xmlParser.js";
 //#region src/docx/fieldParser.ts
 /**
 * All known field types from OOXML specification
@@ -164,10 +164,8 @@ function parseSimpleField(node, styles, theme) {
 		fieldType: parseFieldType(instruction),
 		content: []
 	};
-	const fldLock = getAttribute(node, "w", "fldLock");
-	if (fldLock === "1" || fldLock === "true") field.fldLock = true;
-	const dirty = getAttribute(node, "w", "dirty");
-	if (dirty === "1" || dirty === "true") field.dirty = true;
+	if (parseOnOffAttribute(node, "w", "fldLock") === true) field.fldLock = true;
+	if (parseOnOffAttribute(node, "w", "dirty") === true) field.dirty = true;
 	const children = findChildren(node, "w", "r");
 	for (const child of children) {
 		const run = parseRun(child, styles, theme);

package/dist/docx/footnoteParser.d.ts

@@ -1,6 +1,7 @@
 import { document_d_exports } from "../types/document.js";
 import { NumberingMap } from "./numberingParser.js";
 import { StyleMap } from "./styleParser.js";
+import { ParseContext } from "./parseContext.js";
 import { parseEndnoteProperties, parseFootnoteProperties } from "./notePropertiesParser.js";
 //#region src/docx/footnoteParser.d.ts
 /**
@@ -52,7 +53,7 @@ type EndnoteMap = {
  * @param media - Media files for images
  * @returns FootnoteMap with all footnotes
  */
-declare function parseFootnotes(footnotesXml: string | null, styles?: StyleMap | null, theme?: document_d_exports.Theme | null, numbering?: NumberingMap | null, rels?: document_d_exports.RelationshipMap | null, media?: Map<string, document_d_exports.MediaFile> | null): FootnoteMap;
+declare function parseFootnotes(footnotesXml: string | null, styles?: StyleMap | null, theme?: document_d_exports.Theme | null, numbering?: NumberingMap | null, rels?: document_d_exports.RelationshipMap | null, media?: Map<string, document_d_exports.MediaFile> | null, context?: ParseContext): FootnoteMap;
 /**
  * Parse endnotes.xml
  *
@@ -64,7 +65,7 @@ declare function parseFootnotes(footnotesXml: string | null, styles?: StyleMap |
  * @param media - Media files for images
  * @returns EndnoteMap with all endnotes
  */
-declare function parseEndnotes(endnotesXml: string | null, styles?: StyleMap | null, theme?: document_d_exports.Theme | null, numbering?: NumberingMap | null, rels?: document_d_exports.RelationshipMap | null, media?: Map<string, document_d_exports.MediaFile> | null): EndnoteMap;
+declare function parseEndnotes(endnotesXml: string | null, styles?: StyleMap | null, theme?: document_d_exports.Theme | null, numbering?: NumberingMap | null, rels?: document_d_exports.RelationshipMap | null, media?: Map<string, document_d_exports.MediaFile> | null, context?: ParseContext): EndnoteMap;
 /**
  * Get plain text content of a footnote.
  *

package/dist/docx/footnoteParser.js

@@ -4,6 +4,7 @@ import { parseParagraph } from "./paragraphParser.js";
 import { parseSdtProperties } from "./sdtProperties.js";
 import { parseTable } from "./tableParser.js";
 import { findChild, findChildren, getAttributes, getChildElements, getLocalName, parseXml } from "./xmlParser.js";
+import { PARSE_WARNING_CODES } from "@stll/docx-core/model";
 //#region src/docx/footnoteParser.ts
 /**
 * Parse note type attribute
@@ -69,7 +70,7 @@ function parseFootnote(element, styles, theme, numbering, rels, media) {
 * @param media - Media files for images
 * @returns FootnoteMap with all footnotes
 */
-function parseFootnotes(footnotesXml, styles = null, theme = null, numbering = null, rels = null, media = null) {
+function parseFootnotes(footnotesXml, styles = null, theme = null, numbering = null, rels = null, media = null, context) {
 	const byId = /* @__PURE__ */ new Map();
 	const footnotes = [];
 	if (!footnotesXml) return createFootnoteMap(byId, footnotes);
@@ -78,7 +79,14 @@ function parseFootnotes(footnotesXml, styles = null, theme = null, numbering = n
 	const footnoteElements = findChildren(rootElement, "w", "footnote");
 	for (const fnEl of footnoteElements) {
 		const footnote = parseFootnote(fnEl, styles, theme, numbering, rels, media);
-		if (byId.has(footnote.id)) continue;
+		if (byId.has(footnote.id)) {
+			context?.warn({
+				code: PARSE_WARNING_CODES.duplicateNoteId,
+				element: "w:footnote",
+				at: `w:id ${String(footnote.id)}`
+			});
+			continue;
+		}
 		byId.set(footnote.id, footnote);
 		footnotes.push(footnote);
 	}
@@ -130,7 +138,7 @@ function parseEndnote(element, styles, theme, numbering, rels, media) {
 * @param media - Media files for images
 * @returns EndnoteMap with all endnotes
 */
-function parseEndnotes(endnotesXml, styles = null, theme = null, numbering = null, rels = null, media = null) {
+function parseEndnotes(endnotesXml, styles = null, theme = null, numbering = null, rels = null, media = null, context) {
 	const byId = /* @__PURE__ */ new Map();
 	const endnotes = [];
 	if (!endnotesXml) return createEndnoteMap(byId, endnotes);
@@ -139,7 +147,14 @@ function parseEndnotes(endnotesXml, styles = null, theme = null, numbering = nul
 	const endnoteElements = findChildren(rootElement, "w", "endnote");
 	for (const enEl of endnoteElements) {
 		const endnote = parseEndnote(enEl, styles, theme, numbering, rels, media);
-		if (byId.has(endnote.id)) continue;
+		if (byId.has(endnote.id)) {
+			context?.warn({
+				code: PARSE_WARNING_CODES.duplicateNoteId,
+				element: "w:endnote",
+				at: `w:id ${String(endnote.id)}`
+			});
+			continue;
+		}
 		byId.set(endnote.id, endnote);
 		endnotes.push(endnote);
 	}

package/dist/docx/groupDrawingParser.js

@@ -1,6 +1,7 @@
 import { emuToPixels } from "../utils/units.js";
 import { parseImage, resolveImageData } from "./imageParser.js";
 import { findAllDeep, findChildByLocalName, findChildrenByLocalName, getAttribute, getChildElements, getLocalName, getTextContent, parseNumericAttribute } from "./xmlParser.js";
+import { escapeXmlAttribute, escapeXmlText } from "@stll/docx-core";
 //#region src/docx/groupDrawingParser.ts
 const HEX_COLOR = /^[0-9A-Fa-f]{6}$/u;
 const DEFAULT_TEXT_COLOR = "000000";
@@ -12,7 +13,6 @@ const CROP_SCALE = 1e5;
 const MAX_PATH_COMMANDS = 1e4;
 const MAX_TEXT_CHARACTERS = 2e4;
 const MAX_SVG_CHARACTERS = 1e6;
-const escapeXml = (value) => value.replaceAll("&", "&amp;").replaceAll("<", "&lt;").replaceAll(">", "&gt;").replaceAll("\"", "&quot;").replaceAll("'", "&apos;");
 const numericAttr = (element, name) => {
 	const direct = parseNumericAttribute(element, null, name);
 	if (direct !== void 0) return direct;
@@ -108,7 +108,7 @@ const renderTextBox = (wsp) => {
 	const color = colorFrom(findAllDeep(wsp, "w", "color").at(0) ?? null, DEFAULT_TEXT_COLOR);
 	const fontSize = halfPoints * HALF_POINT_TO_EMU;
 	const maxCharacters = Math.max(1, Math.floor(width / (fontSize * .38)));
-	const lines = paragraphs.flatMap((paragraph) => wrapLine(getTextContent(paragraph).slice(0, MAX_TEXT_CHARACTERS), maxCharacters)).map(escapeXml);
+	const lines = paragraphs.flatMap((paragraph) => wrapLine(getTextContent(paragraph).slice(0, MAX_TEXT_CHARACTERS), maxCharacters)).map(escapeXmlText);
 	if (lines.length === 0) return "";
 	const lineHeight = fontSize * 1.15;
 	const svgFontSize = 1e3;
@@ -121,7 +121,7 @@ const renderPicture = (picture, index, rels, media) => {
 	if (width <= 0 || height <= 0) return "";
 	const blipFill = findChildByLocalName(picture, "blipFill");
 	const blip = findChildByLocalName(blipFill, "blip");
-	const { src } = resolveImageData(getAttribute(blip, "r", "embed") ?? getAttribute(blip, "r", "link") ?? "", rels, media);
+	const { src } = resolveImageData(getAttribute(blip, "r", "embed") ?? getAttribute(blip, "r", "link") ?? void 0, rels, media);
 	if (!src) return "";
 	const sourceRect = findChildByLocalName(blipFill, "srcRect");
 	const left = Math.max(0, numericAttr(sourceRect, "l")) / CROP_SCALE;
@@ -131,7 +131,7 @@ const renderPicture = (picture, index, rels, media) => {
 	const visibleWidth = 1 - left - right;
 	const visibleHeight = 1 - top - bottom;
 	if (visibleWidth <= 0 || visibleHeight <= 0) return "";
-	const image = `<image x="${x - width * left / visibleWidth}" y="${y - height * top / visibleHeight}" width="${width / visibleWidth}" height="${height / visibleHeight}" href="${escapeXml(src)}" preserveAspectRatio="none"/>`;
+	const image = `<image x="${x - width * left / visibleWidth}" y="${y - height * top / visibleHeight}" width="${width / visibleWidth}" height="${height / visibleHeight}" href="${escapeXmlAttribute(src)}" preserveAspectRatio="none"/>`;
 	if (left === 0 && top === 0 && right === 0 && bottom === 0) return image;
 	const clipId = `group-picture-${index}`;
 	return `<defs><clipPath id="${clipId}"><rect x="${x}" y="${y}" width="${width}" height="${height}"/></clipPath></defs><g clip-path="url(#${clipId})">${image}</g>`;

package/dist/docx/headerFooterRefParser.d.ts

@@ -1,4 +1,5 @@
 import { document_d_exports } from "../types/document.js";
+import { ParseContext } from "./parseContext.js";
 import { XmlElement } from "./xmlParser.js";
 //#region src/docx/headerFooterRefParser.d.ts
 /**
@@ -11,15 +12,15 @@ import { XmlElement } from "./xmlParser.js";
  * header/footer reference types has to read them through here, or a
  * normalisation at this boundary looks like a lost reference downstream.
  */
-declare function parseHeaderFooterType(typeAttr: string | null): document_d_exports.HeaderFooterType;
+declare function parseHeaderFooterType(typeAttr: string | null, context?: ParseContext): document_d_exports.HeaderFooterType;
 /**
  * Parse a header reference from sectPr (w:headerReference)
  */
-declare function parseHeaderReference(element: XmlElement): document_d_exports.HeaderReference;
+declare function parseHeaderReference(element: XmlElement, context?: ParseContext): document_d_exports.HeaderReference | null;
 /**
  * Parse a footer reference from sectPr (w:footerReference)
  */
-declare function parseFooterReference(element: XmlElement): document_d_exports.FooterReference;
+declare function parseFooterReference(element: XmlElement, context?: ParseContext): document_d_exports.FooterReference | null;
 /**
  * Parse all header references from a sectPr element
  */

package/dist/docx/headerFooterRefParser.js

@@ -1,6 +1,14 @@
 import { findChildren, getAttribute } from "./xmlParser.js";
+import { PARSE_WARNING_CODES } from "@stll/docx-core/model";
 //#region src/docx/headerFooterRefParser.ts
 /**
+* Header/Footer Reference Parser
+*
+* Parses header/footer references (w:headerReference, w:footerReference) that
+* appear in section properties. Extracted from headerFooterParser to break the
+* circular dependency: headerFooterParser -> paragraphParser -> sectionParser -> headerFooterParser.
+*/
+/**
 * Read a `w:type` attribute as one of ECMA-376's three `ST_HdrFtr` values.
 *
 * The enumeration is `even`, `default` and `first` (17.18.36); `default` is
@@ -10,32 +18,48 @@ import { findChildren, getAttribute } from "./xmlParser.js";
 * header/footer reference types has to read them through here, or a
 * normalisation at this boundary looks like a lost reference downstream.
 */
-function parseHeaderFooterType(typeAttr) {
+function parseHeaderFooterType(typeAttr, context) {
 	switch (typeAttr) {
 		case "first": return "first";
 		case "even": return "even";
-		default: return "default";
+		case "default":
+		case null: return "default";
+		default:
+			context?.warn({
+				code: PARSE_WARNING_CODES.headerFooterTypeOutsideEnum,
+				value: typeAttr,
+				element: "w:type"
+			});
+			return "default";
 	}
 }
-function parseHeaderFooterReference(element) {
-	const typeAttr = getAttribute(element, "w", "type");
-	const rId = getAttribute(element, "r", "id") ?? "";
+/**
+* A reference with no `r:id` names no part, so it is not a reference.
+*
+* Coercing the missing attribute to `""` used to put the empty string in the
+* model, where it became a part-map key on one side and an `r:id=""` the
+* schema rejects on the other. Null here keeps the reference out of the model
+* entirely, which is what the source said.
+*/
+function parseHeaderFooterReference(element, context) {
+	const rId = getAttribute(element, "r", "id");
+	if (rId === null || rId.length === 0) return null;
 	return {
-		type: parseHeaderFooterType(typeAttr),
+		type: parseHeaderFooterType(getAttribute(element, "w", "type"), context?.scoped({ at: `r:id "${rId}"` })),
 		rId
 	};
 }
 /**
 * Parse a header reference from sectPr (w:headerReference)
 */
-function parseHeaderReference(element) {
-	return parseHeaderFooterReference(element);
+function parseHeaderReference(element, context) {
+	return parseHeaderFooterReference(element, context);
 }
 /**
 * Parse a footer reference from sectPr (w:footerReference)
 */
-function parseFooterReference(element) {
-	return parseHeaderFooterReference(element);
+function parseFooterReference(element, context) {
+	return parseHeaderFooterReference(element, context);
 }
 /**
 * Parse all header references from a sectPr element
@@ -43,7 +67,10 @@ function parseFooterReference(element) {
 function parseHeaderReferences(sectPr) {
 	const refs = [];
 	const headerRefElements = findChildren(sectPr, "w", "headerReference");
-	for (const el of headerRefElements) refs.push(parseHeaderReference(el));
+	for (const el of headerRefElements) {
+		const ref = parseHeaderReference(el);
+		if (ref) refs.push(ref);
+	}
 	return refs;
 }
 /**
@@ -52,7 +79,10 @@ function parseHeaderReferences(sectPr) {
 function parseFooterReferences(sectPr) {
 	const refs = [];
 	const footerRefElements = findChildren(sectPr, "w", "footerReference");
-	for (const el of footerRefElements) refs.push(parseFooterReference(el));
+	for (const el of footerRefElements) {
+		const ref = parseFooterReference(el);
+		if (ref) refs.push(ref);
+	}
 	return refs;
 }
 //#endregion

package/dist/docx/headerFooterReferenceNormalization.d.ts

@@ -1,5 +1,8 @@
 import { document_d_exports } from "../types/document.js";
 //#region src/docx/headerFooterReferenceNormalization.d.ts
+/** The codes this normalisation is reported under, owned here, not at the caller. */
+declare const DANGLING_HEADER_REFERENCE_WARNING: "dangling-header-reference";
+declare const DANGLING_FOOTER_REFERENCE_WARNING: "dangling-footer-reference";
 type NormalizeHeaderFooterReferencesInput = {
   documentBody: document_d_exports.DocumentBody;
   headers?: Map<string, document_d_exports.HeaderFooter>;
@@ -11,4 +14,4 @@ type NormalizeHeaderFooterReferencesResult = {
 };
 declare const normalizeHeaderFooterReferences: ({ documentBody, headers, footers }: NormalizeHeaderFooterReferencesInput) => NormalizeHeaderFooterReferencesResult;
 //#endregion
-export { normalizeHeaderFooterReferences };
+export { DANGLING_FOOTER_REFERENCE_WARNING, DANGLING_HEADER_REFERENCE_WARNING, normalizeHeaderFooterReferences };

package/dist/docx/headerFooterReferenceNormalization.js

@@ -1,4 +1,8 @@
+import { PARSE_WARNING_CODES } from "@stll/docx-core/model";
 //#region src/docx/headerFooterReferenceNormalization.ts
+/** The codes this normalisation is reported under, owned here, not at the caller. */
+const DANGLING_HEADER_REFERENCE_WARNING = PARSE_WARNING_CODES.danglingHeaderReference;
+const DANGLING_FOOTER_REFERENCE_WARNING = PARSE_WARNING_CODES.danglingFooterReference;
 const normalizeHeaderFooterReferences = ({ documentBody, headers, footers }) => {
 	const seenSectionProperties = /* @__PURE__ */ new Set();
 	let removedDanglingHeaderReferences = 0;
@@ -61,4 +65,4 @@ const removeDanglingReferences = (references, validParts) => {
 	};
 };
 //#endregion
-export { normalizeHeaderFooterReferences };
+export { DANGLING_FOOTER_REFERENCE_WARNING, DANGLING_HEADER_REFERENCE_WARNING, normalizeHeaderFooterReferences };