@stll/folio-core 0.59.0 → 0.61.0

package/dist/controller/layoutPipeline.js

@@ -42,7 +42,7 @@ const formatEndnoteTexts = (numbers, formatNumber) => {
 	for (const [id, displayNumber] of numbers) texts.set(id, formatNumber(displayNumber));
 	return texts;
 };
-const sameMeasureInputs = (left, right) => left.styles === right.styles && left.theme === right.theme && left.defaultTabStop === right.defaultTabStop && left.pageContentHeight === right.pageContentHeight && left.fontSet === right.fontSet && left.fontAlternates.size === right.fontAlternates.size && [...left.fontAlternates].every(([name, alternate]) => right.fontAlternates.get(name) === alternate) && left.markupView === right.markupView;
+const sameMeasureInputs = (left, right) => left.embeddedFontFamilyMap === right.embeddedFontFamilyMap && left.styles === right.styles && left.theme === right.theme && left.defaultTabStop === right.defaultTabStop && left.pageContentHeight === right.pageContentHeight && left.fontSet === right.fontSet && left.fontAlternates.size === right.fontAlternates.size && [...left.fontAlternates].every(([name, alternate]) => right.fontAlternates.get(name) === alternate) && left.markupView === right.markupView && left.showCommentHighlights === right.showCommentHighlights;
 /**
 * The range an incremental pass re-measures, or `null` for a full measure: the
 * committed measures must come from the same measure inputs, and the range is
@@ -136,7 +136,7 @@ function runLayoutPipeline(deps, state, options = {}) {
 	return result;
 }
 function runLayoutPipelineMeasured(deps, state, options) {
-	const { contentWidth, columns, pageSize, margins, pageGap, showMarginGuides, marginGuideColor, syncCoordinator, headerContent, footerContent, firstPageHeaderContent, firstPageFooterContent, headerContentRId, footerContentRId, firstPageHeaderContentRId, firstPageFooterContentRId, sectionHeaderFooterRefs, theme: _theme, sectionProperties, document, defaultTabStop, mirrorMargins, styles, layout, hfPMs, painter, pagesContainer, session, renderHfFromContentOrPm, renderHeaderFooterContentByRId, readFontSetSignature, buildFootnoteRenderItems, describeInvalidHighlightMarks, pageRenderer = PAGE_RENDERER.legacy, emptyTemplatePreviewEntries: EMPTY_TEMPLATE_PREVIEW_ENTRIES, emptyTemplatePreviewHidden: EMPTY_TEMPLATE_PREVIEW_HIDDEN, markupView } = deps;
+	const { contentWidth, columns, pageSize, margins, pageGap, showMarginGuides, marginGuideColor, syncCoordinator, headerContent, footerContent, firstPageHeaderContent, firstPageFooterContent, headerContentRId, footerContentRId, firstPageHeaderContentRId, firstPageFooterContentRId, sectionHeaderFooterRefs, theme: _theme, sectionProperties, document, defaultTabStop, mirrorMargins, styles, layout, hfPMs, painter, pagesContainer, session, renderHfFromContentOrPm, renderHeaderFooterContentByRId, readFontSetSignature, buildFootnoteRenderItems, describeInvalidHighlightMarks, pageRenderer = PAGE_RENDERER.legacy, emptyTemplatePreviewEntries: EMPTY_TEMPLATE_PREVIEW_ENTRIES, emptyTemplatePreviewHidden: EMPTY_TEMPLATE_PREVIEW_HIDDEN, markupView, showCommentHighlights = true } = deps;
 	let outcome = {};
 	let pendingArtifacts;
 	let pendingTemplatePreview;
@@ -154,16 +154,22 @@ function runLayoutPipelineMeasured(deps, state, options) {
 		let phaseStartedAt = performance.now();
 		const pageContentHeight = pageSize.h - margins.top - margins.bottom;
 		const fontAlternates = buildFontAlternates(document?.package.fontTable);
+		const displayCommentHighlights = showCommentHighlights && markupView !== "no-markup" && markupView !== "original";
+		const embeddedFontFamilies = session.embeddedFontFamilyMap ? { embeddedFontFamilyMap: session.embeddedFontFamilyMap } : {};
 		const measureInputs = {
+			...embeddedFontFamilies,
 			styles,
 			theme: _theme,
 			defaultTabStop,
 			pageContentHeight,
 			fontSet,
 			fontAlternates,
-			markupView
+			markupView,
+			showCommentHighlights: displayCommentHighlights
 		};
 		const flowOpts = {
+			showCommentHighlights: displayCommentHighlights,
+			...embeddedFontFamilies,
 			pageContentHeight,
 			fontAlternates,
 			...styles ? { styles } : {},
@@ -243,11 +249,14 @@ function runLayoutPipelineMeasured(deps, state, options) {
 		const buildHfOptions = (pageCount, fieldInputs) => {
 			const hfMeasureBlocks = (hfBlocks, hfWidth) => measureBlocks(hfBlocks, hfWidth, void 0, void 0, buildHeaderFooterFieldValues(hfBlocks, pageCount, hfClock, fieldInputs), { allowEndTabOverflow: true });
 			return {
+				markupView,
+				showCommentHighlights: displayCommentHighlights,
 				...flowOpts.styles ? { styles: flowOpts.styles } : {},
 				...document?.package.numbering === void 0 ? {} : { numbering: document.package.numbering },
 				...flowOpts.defaultSize === void 0 ? {} : { defaultSize: flowOpts.defaultSize },
 				..._theme !== void 0 ? { theme: _theme } : {},
 				fontAlternates,
+				...embeddedFontFamilies,
 				measureBlocks: hfMeasureBlocks,
 				...defaultTabStop !== void 0 ? { defaultTabStopTwips: defaultTabStop } : {},
 				...flowOpts.lineBreakRules ? { lineBreakRules: flowOpts.lineBreakRules } : {},
@@ -374,12 +383,17 @@ function runLayoutPipelineMeasured(deps, state, options) {
 		outcome.measures = newMeasures;
 		recordPhaseDuration("measure-blocks", phaseStartedAt);
 		const buildNoteOptions = () => {
-			const noteOptions = { measureBlocks };
+			const noteOptions = {
+				measureBlocks,
+				showCommentHighlights: displayCommentHighlights,
+				markupView
+			};
 			if (flowOpts.styles) noteOptions.styles = flowOpts.styles;
 			if (flowOpts.defaultSize !== void 0) noteOptions.defaultSize = flowOpts.defaultSize;
 			if (_theme !== void 0) noteOptions.theme = _theme;
 			if (document?.package.numbering !== void 0) noteOptions.numbering = document.package.numbering;
 			noteOptions.fontAlternates = fontAlternates;
+			if (session.embeddedFontFamilyMap) noteOptions.embeddedFontFamilyMap = session.embeddedFontFamilyMap;
 			if (defaultTabStop !== void 0) noteOptions.defaultTabStopTwips = defaultTabStop;
 			if (flowOpts.lineBreakRules) noteOptions.lineBreakRules = flowOpts.lineBreakRules;
 			if (flowOpts.justificationCompatibility) noteOptions.justificationCompatibility = flowOpts.justificationCompatibility;
@@ -579,6 +593,7 @@ function runLayoutPipelineMeasured(deps, state, options) {
 			outcome.blockLookup = blockLookup;
 			const footnotesByPage = hasFootnotes ? buildFootnoteRenderItems(pageFootnoteMap, footnoteContentMap, document) : void 0;
 			const renderOpts = {
+				...embeddedFontFamilies,
 				pageGap,
 				showShadow: true,
 				blockLookup,

package/dist/controller/layoutSession.d.ts

@@ -1,7 +1,7 @@
 import { document_d_exports } from "../types/document.js";
 import { TemplatePreviewEntry, TemplatePreviewHiddenRange, TemplatePreviewValues } from "../prosemirror/plugins/templatePreviewValues.js";
-import { FlowBlock, Measure } from "../layout-engine/types.js";
 import { DisplayMode } from "../managers/EditorModeManager.js";
+import { FlowBlock, Measure } from "../layout-engine/types.js";
 import { FontAlternates } from "../fonts/fontAlternates.js";
 import { EditorState } from "prosemirror-state";
 //#region src/controller/layoutSession.d.ts
@@ -23,6 +23,7 @@ type LayoutTemplatePreview = {
  * that changes how text measures brings new styles or font alternates.
  */
 type LayoutMeasureInputs = {
+  embeddedFontFamilyMap?: ReadonlyMap<string, string>;
   styles: document_d_exports.StyleDefinitions | null | undefined;
   theme: document_d_exports.Theme | null | undefined;
   defaultTabStop: number | undefined;
@@ -35,8 +36,10 @@ type LayoutMeasureInputs = {
   fontAlternates: FontAlternates;
   /** The review view: each lays out different text, so a measure is reusable only within one. */
   markupView: DisplayMode;
+  showCommentHighlights?: boolean;
 };
 type LayoutSession = {
+  embeddedFontFamilyMap?: ReadonlyMap<string, string>;
   artifacts: LayoutArtifacts | null;
   lastEditorState: EditorState | null;
   lastPmDoc: EditorState["doc"] | null;

package/dist/display-list/build/fontTable.js

@@ -1,6 +1,6 @@
 import { ptToPx } from "../../layout-engine/measure/measureHelpers.js";
 import { getFontMetrics } from "../../layout-engine/measure/measureProvider.js";
-import { resolveFontFamily } from "../../utils/fontResolver.js";
+import { parseFontFamilyList, resolveFontFamily } from "../../utils/fontResolver.js";
 //#region src/display-list/build/fontTable.ts
 const CSS_NORMAL_WEIGHT = 400;
 const CSS_BOLD_WEIGHT = 700;
@@ -12,7 +12,6 @@ const GENERIC_FAMILIES = {
 	fantasy: "fantasy"
 };
 const DEFAULT_GENERIC = "sans-serif";
-const splitFallbackStack = (cssFallback) => cssFallback.split(",").map((entry) => entry.trim().replace(/^["']|["']$/gu, ""));
 /**
 * The category `resolveFontFamily` chose. It is the last CSS generic in the
 * stack it built, which is where `detectFontCategory` puts it; reading it back
@@ -30,8 +29,8 @@ const faceKey = (family, weight, italic) => `${family.toLowerCase()} ${weight} $
 * Package faces keyed by every family name a run can resolve to.
 *
 * `resolveFontFamily` substitutes the scoped `folio-embedded-{nonce}-{name}`
-* family only while the document's embedded-family map is installed; a build
-* that never installed it resolves the same run to the authored name. Both
+* family when the run carries its document-owned embedded-family map; a build
+* without that context resolves the same run to the authored name. Both
 * names therefore key the same face, and the bytes are interned by id so one
 * face's binary appears once however many names reach it.
 */
@@ -64,7 +63,10 @@ var FontTable = class {
 		this.embeddedByKey = indexEmbeddedFonts(embeddedFonts);
 	}
 	intern({ fontFamily, alternateFontFamily, bold, italic, measureStyle }) {
-		const stack = splitFallbackStack(resolveFontFamily(fontFamily, alternateFontFamily).cssFallback);
+		const stack = parseFontFamilyList(resolveFontFamily(fontFamily, {
+			alternateFontName: alternateFontFamily,
+			embeddedFontFamilyMap: measureStyle.embeddedFontFamilyMap
+		}).cssFallback);
 		const family = stack.at(0) ?? fontFamily;
 		const weight = bold ? CSS_BOLD_WEIGHT : CSS_NORMAL_WEIGHT;
 		const isItalic = italic === true;

package/dist/display-list/build/furniture.d.ts

@@ -12,6 +12,7 @@ import { PageComposer } from "./regions.js";
  * caller's `documentFeatures` shows the document had one.
  */
 type PageFurnitureInputs = {
+  embeddedFontFamilyMap?: ReadonlyMap<string, string>;
   /**
    * `w:pgBorders` for the section. Absent: no border is painted, and the gap is
    * reported unless `documentFeatures` says the document has none.

package/dist/display-list/build/furniture.js

@@ -61,6 +61,7 @@ const paintPageFurniture = ({ page, furniture, context }) => {
 	const watermark = !suppliesWatermark(furniture) || selected.watermark === void 0 ? [] : paintWatermark({
 		page,
 		watermark: selected.watermark,
+		...furniture.embeddedFontFamilyMap ? { embeddedFontFamilyMap: furniture.embeddedFontFamilyMap } : {},
 		...furniture.watermarkImageSrc === void 0 ? {} : { imageSrc: furniture.watermarkImageSrc },
 		context
 	});

package/dist/display-list/build/imagePrimitives.js

@@ -1,5 +1,6 @@
 import { parseRotationDegrees } from "../../utils/rotationBoundingBox.js";
 import { sanitizeImageSrc } from "../../utils/sanitizeImageSrc.js";
+import { base64ToBytes } from "../../utils/textEncoding.js";
 import { paintPreview } from "./previewPrimitives.js";
 import { resolveBorderStroke } from "./strokes.js";
 import { UNSUPPORTED_CONSTRUCT } from "./unsupported.js";
@@ -21,12 +22,8 @@ const DATA_URL_PATTERN = /^data:([^;,]*)(;base64)?,(.*)$/su;
 * document the builder reports, not an exception that aborts the build.
 */
 const decodeBase64 = (encoded) => {
-	const decoded = Result.try(() => atob(encoded));
-	if (decoded.isErr()) return;
-	const binary = decoded.value;
-	const bytes = new Uint8Array(binary.length);
-	for (let index = 0; index < binary.length; index += 1) bytes[index] = binary.charCodeAt(index);
-	return bytes;
+	const decoded = Result.try(() => base64ToBytes(encoded));
+	return decoded.isOk() ? decoded.value : void 0;
 };
 const readUint32BE = (bytes, offset) => bytes[offset] << 24 | bytes[offset + 1] << 16 | bytes[offset + 2] << 8 | bytes[offset + 3];
 const readUint16BE = (bytes, offset) => bytes[offset] << 8 | bytes[offset + 1];

package/dist/display-list/build/watermarkPrimitives.d.ts

@@ -4,12 +4,13 @@ import { DisplayPrimitive } from "../types.js";
 import { BuildContext } from "./buildContext.js";
 //#region src/display-list/build/watermarkPrimitives.d.ts
 type WatermarkPaintOptions = {
+  embeddedFontFamilyMap?: ReadonlyMap<string, string> | undefined;
   readonly page: Page;
   readonly watermark: document_d_exports.Watermark;
   /** Resolved `data:` source for a picture watermark; absent means unresolved. */
   readonly imageSrc?: string;
   readonly context: BuildContext;
 };
-declare const paintWatermark: ({ page, watermark, imageSrc, context }: WatermarkPaintOptions) => readonly DisplayPrimitive[];
+declare const paintWatermark: ({ page, watermark, imageSrc, context, embeddedFontFamilyMap }: WatermarkPaintOptions) => readonly DisplayPrimitive[];
 //#endregion
 export { WatermarkPaintOptions, paintWatermark };

package/dist/display-list/build/watermarkPrimitives.js

@@ -29,9 +29,10 @@ const TEXT_DEFAULT_OPACITY = .5;
 const TEXT_DIAGONAL_DEGREES = -45;
 const PICTURE_NATIVE_SCALE = 1;
 const PICTURE_WASHOUT_OPACITY = .18;
-const paintTextWatermark = (watermark, page, context) => {
+const paintTextWatermark = ({ watermark, page, context, embeddedFontFamilyMap }) => {
 	if (watermark.text.length === 0) return [];
 	const style = {
+		...embeddedFontFamilyMap ? { embeddedFontFamilyMap } : {},
 		fontFamily: watermark.font ?? TEXT_DEFAULT_FONT,
 		fontSize: pxToPt(TEXT_FONT_SIZE_PX),
 		bold: true
@@ -129,9 +130,14 @@ const paintPictureWatermark = (watermark, page, imageSrc, context) => {
 		opacity: watermark.washout === false ? 1 : PICTURE_WASHOUT_OPACITY
 	}];
 };
-const paintWatermark = ({ page, watermark, imageSrc, context }) => {
+const paintWatermark = ({ page, watermark, imageSrc, context, embeddedFontFamilyMap }) => {
 	switch (watermark.kind) {
-		case "text": return paintTextWatermark(watermark, page, context);
+		case "text": return paintTextWatermark({
+			watermark,
+			page,
+			context,
+			embeddedFontFamilyMap
+		});
 		case "picture": return paintPictureWatermark(watermark, page, imageSrc, context);
 		default: return [];
 	}

package/dist/display-list/editor/pageRenderer.js

@@ -17,6 +17,7 @@ const isPageRendererName = (value) => value === PAGE_RENDERER.legacy || value ==
 * which header a page carries.
 */
 const displayListFurnitureFrom = (options) => ({
+	...options.embeddedFontFamilyMap ? { embeddedFontFamilyMap: options.embeddedFontFamilyMap } : {},
 	...options.pageBorders === void 0 ? {} : { pageBorders: options.pageBorders },
 	...options.theme === void 0 ? {} : { theme: options.theme },
 	...options.watermark === void 0 ? {} : { watermark: options.watermark },

package/dist/docx/bookmarkIdNormalization.d.ts

@@ -0,0 +1,10 @@
+//#region src/docx/bookmarkIdNormalization.d.ts
+/**
+ * Give repeated bookmark ranges independent identities before editor pairing.
+ * The first range keeps its id. Ends close starts in encounter order, including
+ * overlapping ambiguous ranges. Only colliding ids are repaired: those require
+ * unique XML ids for a valid save. Every other source byte stays verbatim.
+ */
+declare const normalizeBookmarkIdsInXmlParts: (parts: ReadonlyMap<string, string>) => ReadonlyMap<string, string>;
+//#endregion
+export { normalizeBookmarkIdsInXmlParts };

package/dist/docx/bookmarkIdNormalization.js

@@ -0,0 +1,100 @@
+import { scanSelectedStreamingXmlElements } from "./selectedXmlIdentityScan.js";
+import { WORDPROCESSINGML_NAMESPACE_URIS, findAttributeByNamespaceUri, getLocalName, getNamespaceUri } from "./xmlParser.js";
+import { panic } from "better-result";
+//#region src/docx/bookmarkIdNormalization.ts
+const bookmarkIdentity = (element) => {
+	if (!WORDPROCESSINGML_NAMESPACE_URIS.has(getNamespaceUri(element) ?? "")) return null;
+	const name = getLocalName(element.name);
+	if (name !== "bookmarkStart" && name !== "bookmarkEnd") return null;
+	const attribute = findAttributeByNamespaceUri(element, WORDPROCESSINGML_NAMESPACE_URIS, "id");
+	if (!attribute) return null;
+	const id = Number(attribute.value);
+	if (!Number.isSafeInteger(id) || id < 0) return null;
+	return {
+		name,
+		attribute: attribute.name,
+		id
+	};
+};
+/**
+* Give repeated bookmark ranges independent identities before editor pairing.
+* The first range keeps its id. Ends close starts in encounter order, including
+* overlapping ambiguous ranges. Only colliding ids are repaired: those require
+* unique XML ids for a valid save. Every other source byte stays verbatim.
+*/
+const normalizeBookmarkIdsInXmlParts = (parts) => {
+	const candidates = [...parts].filter(([, xml]) => /bookmark(?:Start|End)/u.test(xml));
+	const reserved = /* @__PURE__ */ new Set();
+	const repeated = /* @__PURE__ */ new Set();
+	const markersByPart = /* @__PURE__ */ new Map();
+	for (const [path, xml] of candidates) {
+		const starts = /* @__PURE__ */ new Set();
+		const markers = [];
+		if (scanSelectedStreamingXmlElements(xml, ({ element, attributeValueSpans }) => {
+			const marker = bookmarkIdentity(element);
+			if (!marker) return;
+			const span = attributeValueSpans.get(marker.attribute);
+			if (!span) panic("Missing bookmark identity source span");
+			markers.push({
+				marker,
+				...span
+			});
+			reserved.add(marker.id);
+			if (marker.name === "bookmarkStart") {
+				if (starts.has(marker.id)) repeated.add(path);
+				starts.add(marker.id);
+			}
+		}).status === "unsupported") {
+			repeated.delete(path);
+			continue;
+		}
+		markersByPart.set(path, markers);
+	}
+	if (repeated.size === 0) return parts;
+	let next = 0;
+	const allocate = () => {
+		while (reserved.has(next)) next += 1;
+		const id = next++;
+		reserved.add(id);
+		return id;
+	};
+	const normalized = new Map(parts);
+	for (const [path, xml] of candidates) {
+		if (!repeated.has(path)) continue;
+		const seen = /* @__PURE__ */ new Set();
+		const open = /* @__PURE__ */ new Map();
+		const chunks = [];
+		let cursor = 0;
+		for (const { marker, start, end } of markersByPart.get(path) ?? []) {
+			let id = marker.id;
+			if (marker.name === "bookmarkStart") {
+				if (seen.has(id)) id = allocate();
+				seen.add(marker.id);
+				let queue = open.get(marker.id);
+				if (!queue) {
+					queue = {
+						ids: [],
+						cursor: 0
+					};
+					open.set(marker.id, queue);
+				}
+				queue.ids.push(id);
+			} else {
+				const queue = open.get(marker.id);
+				const paired = queue?.ids.at(queue.cursor);
+				if (paired !== void 0 && queue) {
+					id = paired;
+					queue.cursor += 1;
+				}
+			}
+			if (id === marker.id) continue;
+			chunks.push(xml.slice(cursor, start), String(id));
+			cursor = end;
+		}
+		chunks.push(xml.slice(cursor));
+		normalized.set(path, chunks.join(""));
+	}
+	return normalized;
+};
+//#endregion
+export { normalizeBookmarkIdsInXmlParts };

package/dist/docx/bookmarkParser.js

@@ -31,8 +31,10 @@ function createBookmarkMap() {
 * @param bookmark - The bookmark start to add
 */
 function addBookmark(map, bookmark) {
-	map.byId.set(bookmark.id, bookmark);
-	if (bookmark.name) map.byName.set(bookmark.name, bookmark);
+	if (!map.byId.has(bookmark.id)) map.byId.set(bookmark.id, bookmark);
+	if (bookmark.name) {
+		if (!map.byName.has(bookmark.name)) map.byName.set(bookmark.name, bookmark);
+	}
 	map.bookmarks.push(bookmark);
 }
 /**

package/dist/docx/canonicalStoryRepack.js

@@ -1,5 +1,6 @@
 import { withTrackedSectionEndpointRemoval } from "../internal/sectionEndpointResolution.js";
 import { withSectionReferenceResolution } from "../internal/sectionReferenceResolution.js";
+import { decodeText, encodeUtf8 } from "../utils/textEncoding.js";
 import { readDocumentSectionFacts } from "./documentSectionFacts.js";
 import { parseDocx } from "./parser.js";
 import { assertSectionCarriersMatchModel } from "./rezip.js";
@@ -9,7 +10,7 @@ import JSZip from "jszip";
 /** Canonical snapshots own explicit section references; the PM projection does not. */
 const referenceKey = ({ element, type, rId }) => `${element}:${type}:${rId}`;
 const xmlFingerprint = async (xml) => {
-	const digest = await crypto.subtle.digest("SHA-256", new TextEncoder().encode(xml));
+	const digest = await crypto.subtle.digest("SHA-256", encodeUtf8(xml));
 	let fingerprint = "";
 	for (const byte of new Uint8Array(digest)) fingerprint += byte.toString(16).padStart(2, "0");
 	return fingerprint;
@@ -18,7 +19,7 @@ const xmlFingerprint = async (xml) => {
 const repackWithCanonicalStoryRemovals = async ({ document, repack }) => {
 	const originalBuffer = document.originalBuffer;
 	if (!originalBuffer) return repack();
-	const xml = await (await JSZip.loadAsync(originalBuffer)).file("word/document.xml")?.async("text");
+	const xml = await (await JSZip.loadAsync(originalBuffer)).file("word/document.xml")?.async("uint8array").then((bytes) => decodeText(bytes, { ignoreBOM: true }));
 	if (!xml) return repack();
 	const before = readDocumentSectionFacts(xml);
 	const serializedXml = serializeDocument(document);

package/dist/docx/commentTextParagraph.d.ts

@@ -0,0 +1,6 @@
+import { document_d_exports } from "../types/document.js";
+//#region src/docx/commentTextParagraph.d.ts
+/** Plain comment text has no authored paragraph formatting. */
+declare const commentTextParagraph: (text: string) => document_d_exports.Paragraph;
+//#endregion
+export { commentTextParagraph };

package/dist/docx/commentTextParagraph.js

@@ -0,0 +1,15 @@
+//#region src/docx/commentTextParagraph.ts
+/** Plain comment text has no authored paragraph formatting. */
+const commentTextParagraph = (text) => ({
+	type: "paragraph",
+	content: [{
+		type: "run",
+		formatting: {},
+		content: [{
+			type: "text",
+			text
+		}]
+	}]
+});
+//#endregion
+export { commentTextParagraph };

package/dist/docx/conformance.js

@@ -1,3 +1,4 @@
+import { sliceText } from "../utils/sliceText.js";
 import { getLocalName, getNamespacePrefix, parseXmlDocument } from "./xmlParser.js";
 import { DOCX_CONFORMANCE_CLASSES } from "@stll/docx-core/model";
 //#region src/docx/conformance.ts
@@ -62,7 +63,11 @@ function scanRootStartTag(scanTarget) {
 }
 const detectDocxConformanceClass = (documentXml) => {
 	if (documentXml === null) return DOCX_CONFORMANCE_CLASSES.UNKNOWN;
-	const rootStartTag = scanRootStartTag(documentXml.length <= ROOT_START_TAG_SCAN_LIMIT ? documentXml : documentXml.slice(0, ROOT_START_TAG_SCAN_LIMIT));
+	const rootStartTag = scanRootStartTag(documentXml.length <= ROOT_START_TAG_SCAN_LIMIT ? documentXml : sliceText({
+		text: documentXml,
+		units: "codeUnits",
+		end: ROOT_START_TAG_SCAN_LIMIT
+	}));
 	const root = (rootStartTag === void 0 ? null : parseXmlDocument(rootStartTag.endsWith("/>") ? rootStartTag : `${rootStartTag.slice(0, -1)}/>`)) ?? parseXmlDocument(documentXml);
 	return detectDocxConformanceClassFromRoot(root);
 };

package/dist/docx/diagramPreview.js

@@ -1,3 +1,4 @@
+import { decodeText } from "../utils/textEncoding.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
@@ -31,7 +32,7 @@ const partByRelationshipId = ({ rels, media, rId, type, cache }) => {
 	const cached = cache.get(path);
 	if (cached !== void 0) return cached;
 	const file = media.get(path);
-	const parsed = file?.data ? parseXmlDocument(new TextDecoder().decode(file.data)) : null;
+	const parsed = file?.data ? parseXmlDocument(decodeText(new Uint8Array(file.data))) : null;
 	cache.set(path, parsed);
 	return parsed;
 };

package/dist/docx/documentParser.d.ts

@@ -19,6 +19,18 @@ declare function extractTemplateVariables(text: string): string[];
  * @returns Array of unique variable names
  */
 declare function extractAllTemplateVariables(content: document_d_exports.BlockContent[]): string[];
+/**
+ * Build sections from content based on section properties in paragraphs
+ *
+ * In OOXML, sections are delimited by:
+ * 1. w:pPr/w:sectPr within a paragraph (marks end of a section)
+ * 2. w:body/w:sectPr (final section properties)
+ *
+ * @param content - All block content
+ * @param finalSectPr - Final section properties from body
+ * @returns Array of sections
+ */
+declare function buildSections(content: document_d_exports.BlockContent[], finalSectPr: document_d_exports.SectionProperties | undefined): document_d_exports.Section[];
 /**
  * Parse document.xml body content
  *
@@ -67,4 +79,4 @@ declare function getSectionCount(body: document_d_exports.DocumentBody): number;
  */
 declare function hasTemplateVariables(body: document_d_exports.DocumentBody): boolean;
 //#endregion
-export { extractAllTemplateVariables, extractTemplateVariables, getAllParagraphs, getAllTables, getParagraphCount, getSectionCount, hasTemplateVariables, parseDocumentBody, parseDocumentBodyTree };
+export { buildSections, extractAllTemplateVariables, extractTemplateVariables, getAllParagraphs, getAllTables, getParagraphCount, getSectionCount, hasTemplateVariables, parseDocumentBody, parseDocumentBodyTree };

package/dist/docx/documentParser.js

@@ -257,4 +257,4 @@ function hasTemplateVariables(body) {
 	return extractAllTemplateVariables(body.content).length > 0;
 }
 //#endregion
-export { extractAllTemplateVariables, extractTemplateVariables, getAllParagraphs, getAllTables, getParagraphCount, getSectionCount, hasTemplateVariables, parseDocumentBody, parseDocumentBodyTree };
+export { buildSections, extractAllTemplateVariables, extractTemplateVariables, getAllParagraphs, getAllTables, getParagraphCount, getSectionCount, hasTemplateVariables, parseDocumentBody, parseDocumentBodyTree };

package/dist/docx/encryption/compoundFile.d.ts

@@ -1,12 +1,4 @@
 //#region src/docx/encryption/compoundFile.d.ts
-/**
- * Read-only MS-CFB compound file access for encrypted OOXML packages.
- *
- * Encrypted `.docx` files store `EncryptionInfo` and `EncryptedPackage` streams
- * at the compound-file root. This reader implements only what MS-OFFCRYPTO needs.
- *
- * @see https://learn.microsoft.com/en-us/openspecs/windows_protocols/ms-cfb/
- */
 type CompoundFileReader = {
   readStream: (absolutePath: string) => Uint8Array | null;
 };

package/dist/docx/encryption/compoundFile.js

@@ -1,3 +1,4 @@
+import { decodeText } from "../../utils/textEncoding.js";
 import { joinBytes } from "./cryptoBytes.js";
 //#region src/docx/encryption/compoundFile.ts
 /**
@@ -39,7 +40,7 @@ const u64 = (view, offset) => {
 	return BigInt(view.getUint32(offset + 4, true)) << 32n | low;
 };
 const decodeName = (nameBytes) => {
-	const decoded = new TextDecoder("utf-16le").decode(nameBytes);
+	const decoded = decodeText(nameBytes, { encoding: "utf-16le" });
 	let end = decoded.length;
 	while (end > 0 && decoded.charCodeAt(end - 1) === 0) end--;
 	return decoded.slice(0, end);

package/dist/docx/encryption/cryptoBytes.d.ts

@@ -2,10 +2,10 @@
 /** Byte helpers shared by MS-CFB and MS-OFFCRYPTO agile decryption. */
 declare const joinBytes: (parts: readonly Uint8Array[]) => Uint8Array;
 declare const writeUint32Le: (value: number) => Uint8Array;
-declare const passwordToUtf16Le: (password: string) => Uint8Array;
+declare const passwordToUtf16Le: (text: string) => Uint8Array<ArrayBuffer>;
 declare const bytesEqual: (left: Uint8Array, right: Uint8Array) => boolean;
 declare const toArrayBuffer: (bytes: Uint8Array) => ArrayBuffer;
-declare const decodeBase64: (encoded: string) => Uint8Array;
+declare const decodeBase64: (text: string) => Uint8Array<ArrayBuffer>;
 declare const padToBlock: (bytes: Uint8Array, blockSize: number, padByte?: number) => Uint8Array;
 //#endregion
 export { bytesEqual, decodeBase64, joinBytes, padToBlock, passwordToUtf16Le, toArrayBuffer, writeUint32Le };

package/dist/docx/encryption/cryptoBytes.js

@@ -1,3 +1,4 @@
+import { base64ToBytes, encodeUtf16LE } from "../../utils/textEncoding.js";
 //#region src/docx/encryption/cryptoBytes.ts
 /** Byte helpers shared by MS-CFB and MS-OFFCRYPTO agile decryption. */
 const joinBytes = (parts) => {
@@ -19,15 +20,7 @@ const writeUint32Le = (value) => {
 	out[3] = value >>> 24 & 255;
 	return out;
 };
-const passwordToUtf16Le = (password) => {
-	const out = new Uint8Array(password.length * 2);
-	for (let i = 0; i < password.length; i++) {
-		const code = password.charCodeAt(i);
-		out[i * 2] = code & 255;
-		out[i * 2 + 1] = code >> 8;
-	}
-	return out;
-};
+const passwordToUtf16Le = encodeUtf16LE;
 const bytesEqual = (left, right) => {
 	if (left.length !== right.length) return false;
 	let mismatch = 0;
@@ -38,15 +31,7 @@ const toArrayBuffer = (bytes) => {
 	if (bytes.buffer instanceof ArrayBuffer && bytes.byteOffset === 0 && bytes.byteLength === bytes.buffer.byteLength) return bytes.buffer;
 	return bytes.buffer.slice(bytes.byteOffset, bytes.byteOffset + bytes.byteLength);
 };
-const decodeBase64 = (encoded) => {
-	if (typeof atob === "function") {
-		const binary = atob(encoded);
-		const out = new Uint8Array(binary.length);
-		for (let i = 0; i < binary.length; i++) out[i] = binary.charCodeAt(i);
-		return out;
-	}
-	return new Uint8Array(Buffer.from(encoded, "base64"));
-};
+const decodeBase64 = base64ToBytes;
 const padToBlock = (bytes, blockSize, padByte = 54) => {
 	if (bytes.length >= blockSize) return bytes.subarray(0, blockSize);
 	const out = new Uint8Array(blockSize);

package/dist/docx/encryption/encryptionInfo.d.ts

@@ -1,9 +1,4 @@
 //#region src/docx/encryption/encryptionInfo.d.ts
-/**
- * Parse the EncryptionInfo stream for Agile Encryption (Office 2010+).
- *
- * @see https://learn.microsoft.com/en-us/openspecs/office_file_formats/ms-offcrypto/87020a34-e73f-4139-99bc-bbdf6cf6fa55
- */
 type AgileKeyMaterial = {
   packageSalt: Uint8Array;
   packageKeyBits: number;

package/dist/docx/encryption/encryptionInfo.js

@@ -1,3 +1,4 @@
+import { decodeText } from "../../utils/textEncoding.js";
 import { decodeBase64 } from "./cryptoBytes.js";
 import { DOCX_ENCRYPTION_ERROR_CODES, DocxEncryptionError } from "./errors.js";
 //#region src/docx/encryption/encryptionInfo.ts
@@ -85,7 +86,7 @@ const requireSpinCount = (attrs, element) => {
 	return spinCount;
 };
 const parseAgileXml = (xmlBytes) => {
-	const xml = new TextDecoder("utf-8").decode(xmlBytes);
+	const xml = decodeText(xmlBytes, { encoding: "utf-8" });
 	const keyDataTag = findElement(xml, "keyData");
 	const integrityTag = findElement(xml, "dataIntegrity");
 	const encryptedKeyTag = findElement(xml, "encryptedKey");

package/dist/docx/ensureParaIds.js

@@ -1,4 +1,5 @@
 import { deterministicHexId } from "../utils/hexId.js";
+import { encodeUtf8 } from "../utils/textEncoding.js";
 import { resolvePackageRelationshipTarget } from "./packageParts.js";
 import { spliceXml } from "./selectiveXmlPatch.js";
 import { loadDocxArchive } from "./server/boundedArchive.js";
@@ -390,7 +391,7 @@ const ensureParaIdsInternal = async (docx, options) => {
 		writeZipPart({
 			zip,
 			path: partPath,
-			data: content,
+			data: encodeUtf8(content),
 			options: {
 				compression: "DEFLATE",
 				compressionOptions: { level: 6 },

package/dist/docx/groupDrawingParser.js

@@ -1,3 +1,4 @@
+import { sliceText } from "../utils/sliceText.js";
 import { emuToPixels } from "../utils/units.js";
 import { collectGroupTextBoxes, readTransform } from "./drawingGroupChildren.js";
 import { parseImage, resolveImageData } from "./imageParser.js";
@@ -142,7 +143,11 @@ const renderTextBox = (wsp, unitScale) => {
 	const color = colorFrom(findAllDeep(wsp, "w", "color").at(0) ?? null, DEFAULT_TEXT_COLOR);
 	const fontSize = halfPoints * HALF_POINT_TO_EMU / unitScale;
 	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(escapeXmlText);
+	const lines = paragraphs.flatMap((paragraph) => wrapLine(sliceText({
+		text: getTextContent(paragraph),
+		units: "codeUnits",
+		end: MAX_TEXT_CHARACTERS
+	}), maxCharacters)).map(escapeXmlText);
 	if (lines.length === 0) return "";
 	const lineHeight = fontSize * 1.15;
 	const svgFontSize = 1e3;