@stll/folio-core 0.52.0 → 0.54.0

package/dist/docx/paragraphParser.js

@@ -2,6 +2,7 @@ import { isValidHexId } from "../utils/hexId.js";
 import { attributeRemainder } from "./attributeRemainder.js";
 import { parseBookmarkEnd as parseBookmarkEnd$1, parseBookmarkStart as parseBookmarkStart$1 } from "./bookmarkParser.js";
 import { CAPTURE, dispatchChildrenWithContext, ownedElsewhere, transitionalNamespaceOf, withPreservedChildren } from "./containerChildren.js";
+import { resolveDefaultParagraphStyle } from "./defaultParagraphStyle.js";
 import { parseFieldType } from "./fieldParser.js";
 import { fieldStateOf, parseFieldState } from "./fieldState.js";
 import { HYPERLINK_CHILD_HANDLERS, parseHyperlink as parseHyperlink$1, parseHyperlinkShell } from "./hyperlinkParser.js";
@@ -578,8 +579,8 @@ const hasRunPayload = ({ run, runElement, rels, media }) => {
 * the assembled field, and one that reaches the editor's inline converter on
 * its own is dropped. So a `w:fldChar` whose field this paragraph never closed,
 * or one with no `begin` before it, loses the editor round trip — and the
-* model holds nothing of its `w:ffData` either, so the save loses a legacy form
-* field's name, macros, help text and checkbox state as well.
+* orphan field has no owning `ComplexField` to hold its `w:ffData`, so dropping
+* this run would also lose its name, macros, help text and checkbox state.
 *
 * `preservedXml` is the run's own capture member: the editor carries it as an
 * opaque atom and the serializer replays the source bytes, `w:ffData` included.
@@ -610,9 +611,11 @@ const LEGACY_FORM_CHECKBOX_GLYPHS = {
 	unchecked: "☐"
 };
 const LEGACY_FORM_CHECKBOX_INSTRUCTION = "FORMCHECKBOX";
-function getLegacyFormCheckboxDisplay(runElement) {
-	const fieldChar = findChild(runElement, "w", "fldChar");
-	const fieldData = fieldChar ? findChild(fieldChar, "w", "ffData") : null;
+const beginFieldCharOf = (runElement) => findChildrenByNamespaceUri(runElement, WORDPROCESSINGML_NAMESPACE_URIS, "fldChar").find((fieldChar) => {
+	const charType = getAttribute(fieldChar, "w", "fldCharType");
+	return charType !== "separate" && charType !== "end";
+});
+function getLegacyFormCheckboxDisplay(fieldData) {
 	const checkBox = fieldData ? findChild(fieldData, "w", "checkBox") : null;
 	if (!checkBox) return;
 	const checked = findChild(checkBox, "w", "checked");
@@ -676,6 +679,30 @@ const FIELD_INLINE_WRAPPER_HANDLERS = {
 	pPr: CAPTURE
 };
 /**
+* Whether the walk is currently inside some open field's code region — its
+* own `begin` has been read but not yet its `separate` — at any nesting
+* depth. `codeDepthStack` tracks every open field regardless of whether it
+* assembles into a `ComplexField`, so this answers the question for content,
+* such as a `w:fldSimple`, that a `ComplexField`'s own code/result split
+* cannot carry.
+*/
+const isHiddenByOpenFieldCode = (scan) => scan.codeDepthStack.some((pastSeparator) => !pastSeparator);
+/**
+* A run captured for a position this field's `fieldResult` (or `fieldCode`)
+* must hold — the source ordering a save has to reproduce puts it there —
+* whose actual content sits inside a deeper, still-open field's own
+* instruction region and so must show nothing. Every non-`w:rPr` child is
+* captured through `preserveRunChild`, the same verbatim-with-no-text capture
+* an unrecognised run child already gets, rather than kept as the `text`/
+* `instrText` the parse produced — a `w:t` this walk read as authored result
+* text is exactly what must not read back out of `fieldResult`.
+*/
+const hiddenRunCaptureOf = (runElement, formatting) => ({
+	type: "run",
+	content: getChildElements(runElement).filter((child) => getLocalName(child.name) !== "rPr").map((child) => preserveRunChild(child)),
+	...formatting && Object.keys(formatting).length > 0 ? { formatting } : {}
+});
+/**
 * The container children a complex field's result may hold and still be
 * assembled into a `ComplexField`: its runs and the zero-width markers read
 * beside them. Any other child is inline content (a hyperlink, a simple field,
@@ -713,25 +740,31 @@ const isWordprocessingChild = (child) => {
 };
 /**
 * The `w:r` children whose `begin` opens a complex field that this container's
-* walk closes.
+* walk closes as its own top-level `ComplexField`.
 *
 * A field is assembled from the runs between its `begin` and its `end`, so it
 * exists only when both are runs of this container with nothing between them
 * that a `ComplexField` cannot hold. A TOC opened here and closed in a later
-* paragraph, a field whose result holds a `w:hyperlink` (a TOC's `\h` entries),
-* and an outer field whose result holds a nested field are none of these; their
-* runs are ordinary content, read in source order alongside the siblings
-* between them. Mirrors the `w:r` handler run by run: a run with a `begin`
-* opens a field, and a run with an `end` closes the one open.
+* paragraph, and a field whose result holds a `w:hyperlink` (a TOC's `\h`
+* entries), are neither of those; their runs are ordinary content, read in
+* source order alongside the siblings between them.
+*
+* A `begin`/`end` pair that closes while another `begin` is still open is
+* nested inside that other field's code or result — its own runs stay with
+* the outer field's, carried in `fieldCode`/`fieldResult` rather than
+* assembled into a `ComplexField` of their own, which is why only the
+* outermost pair of a nested run adds its `begin` here. Tracked with a stack
+* rather than one slot so a field nested inside a field nested inside a field
+* closes each level against the right opener instead of the most recent one.
 */
 const assembledFieldBeginsOf = (container) => {
 	const assembled = /* @__PURE__ */ new Set();
-	let open;
+	const open = [];
 	for (const child of getChildElements(container)) {
 		const wordprocessing = isWordprocessingChild(child);
 		const localName = getLocalName(child.name);
 		if (!wordprocessing || localName !== "r") {
-			if (!wordprocessing || !FIELD_TRANSPARENT_CHILDREN.has(localName)) open = void 0;
+			if (!wordprocessing || !FIELD_TRANSPARENT_CHILDREN.has(localName)) open.length = 0;
 			continue;
 		}
 		let hasBegin = false;
@@ -742,10 +775,10 @@ const assembledFieldBeginsOf = (container) => {
 			if (charType === "end") hasEnd = true;
 			else if (charType !== "separate") hasBegin = true;
 		}
-		if (hasBegin) open = child;
-		if (hasEnd && open !== void 0) {
-			assembled.add(open);
-			open = void 0;
+		if (hasBegin) open.push(child);
+		if (hasEnd && open.length > 0) {
+			const closed = open.pop();
+			if (closed !== void 0 && open.length === 0) assembled.add(closed);
 		}
 	}
 	return assembled;
@@ -774,7 +807,8 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 		const commentReferenceId = getCommentReferenceId(runElement);
 		let hasFieldBegin = false;
 		let beginFieldState = {};
-		const beginFallbackDisplay = getLegacyFormCheckboxDisplay(runElement);
+		const beginFieldData = findWordprocessingChild(beginFieldCharOf(runElement), "ffData");
+		const beginFallbackDisplay = getLegacyFormCheckboxDisplay(beginFieldData);
 		let hasFieldSeparate = false;
 		let hasFieldEnd = false;
 		let endOriginalValue;
@@ -788,6 +822,8 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 			if (content.originalValue !== void 0) endOriginalValue = content.originalValue;
 		}
 		else if (content.type === "instrText") instrText += content.text;
+		if (hasFieldBegin) scan.codeDepthStack.push(false);
+		if (hasFieldSeparate && scan.codeDepthStack.length > 0) scan.codeDepthStack[scan.codeDepthStack.length - 1] = true;
 		if (hasFieldBegin) {
 			scan.assembledFieldBegins ??= assembledFieldBeginsOf(scan.container);
 			if (!scan.assembledFieldBegins.has(child)) hasFieldBegin = false;
@@ -804,9 +840,13 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 			scan.complexFieldResultRuns = [];
 			scan.complexFieldOpenRuns = [];
 			scan.complexFieldState = beginFieldState;
+			scan.complexFieldDataXml = beginFieldData ? captureVerbatimXml(beginFieldData) : void 0;
 			scan.complexFieldFallbackDisplay = beginFallbackDisplay;
 			scan.complexFieldFormatting = run.formatting;
+			scan.complexFieldBaseDepth = scan.codeDepthStack.length;
 		}
+		const isNestedField = scan.codeDepthStack.length > scan.complexFieldBaseDepth;
+		if (hasFieldEnd && scan.codeDepthStack.length > 0) scan.codeDepthStack.pop();
 		if (scan.inComplexField) {
 			scan.complexFieldOpenRuns.push(withOrphanFieldCharsPreserved(run, runElement));
 			if (instrText) scan.complexFieldInstr += instrText;
@@ -815,6 +855,10 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 			const fieldCodeContent = [];
 			for (const content of run.content) {
 				if (content.type === "fieldChar") {
+					if (isNestedField) {
+						if (inFieldCodeRegion) fieldCodeContent.push(content);
+						continue;
+					}
 					inFieldCodeRegion = content.charType === "begin";
 					continue;
 				}
@@ -824,11 +868,11 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 				...run,
 				content: fieldCodeContent
 			});
-			if (hasFieldSeparate) scan.afterSeparator = true;
+			if (hasFieldSeparate && !isNestedField) scan.afterSeparator = true;
 			if (scan.afterSeparator && !hasFieldEnd) {
-				if (!hasFieldSeparate) scan.complexFieldResultRuns.push(run);
+				if (!hasFieldSeparate) scan.complexFieldResultRuns.push(isHiddenByOpenFieldCode(scan) ? hiddenRunCaptureOf(runElement, run.formatting) : run);
 			}
-			if (hasFieldEnd) {
+			if (hasFieldEnd && !isNestedField) {
 				let resultRuns = scan.complexFieldResultRuns;
 				let resultIsFallback = false;
 				if (resultRuns.length === 0 && scan.complexFieldFallbackDisplay !== void 0 && isLegacyFormCheckboxInstruction(scan.complexFieldInstr)) {
@@ -850,6 +894,7 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 					fieldResult: resultRuns,
 					...scan.complexFieldState
 				};
+				if (scan.complexFieldDataXml !== void 0) complexField.formFieldDataXml = scan.complexFieldDataXml;
 				if (resultIsFallback) complexField.fieldResultIsFallback = true;
 				if (scan.complexFieldFormatting) complexField.formatting = scan.complexFieldFormatting;
 				contents.push(complexField);
@@ -865,6 +910,17 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 				type: "commentReference",
 				id: commentReferenceId
 			});
+		} else if (isHiddenByOpenFieldCode(scan)) {
+			if (hasRunPayload({
+				run,
+				runElement,
+				rels,
+				media
+			})) contents.push({
+				type: "preservedInline",
+				xml: captureVerbatimXml(runElement),
+				text: ""
+			});
 		} else if (hasRunPayload({
 			run,
 			runElement,
@@ -881,8 +937,17 @@ const PARAGRAPH_CONTENT_HANDLERS = {
 	bookmarkEnd: (child, { contents }) => {
 		contents.push(parseBookmarkEnd(child));
 	},
-	fldSimple: (child, { contents, styles, theme, rels, media, previews, inScopeXmlns }) => {
-		contents.push(parseSimpleField(child, styles, theme, rels, media, inScopeXmlns, previews));
+	fldSimple: (child, { contents, styles, theme, rels, media, previews, inScopeXmlns, trackedContext, scan }) => {
+		const fieldElement = trackedContext === "deletion" ? normalizeDeletionContentElement(child) : child;
+		if (isHiddenByOpenFieldCode(scan)) {
+			contents.push({
+				type: "preservedInline",
+				xml: captureVerbatimXml(fieldElement),
+				text: ""
+			});
+			return;
+		}
+		contents.push(parseSimpleField(fieldElement, styles, theme, rels, media, inScopeXmlns, previews));
 	},
 	pPr: PARAGRAPH_PROPERTIES_OWNER,
 	customXml: paragraphInlineWrapper("customXml"),
@@ -1005,10 +1070,13 @@ function parseParagraphContents(paraElement, styles, theme, _numbering, rels, me
 		complexFieldOpenRuns: [],
 		afterSeparator: false,
 		complexFieldState: {},
+		complexFieldDataXml: void 0,
 		complexFieldFallbackDisplay: void 0,
 		complexFieldFormatting: void 0,
 		container: paraElement,
-		assembledFieldBegins: void 0
+		assembledFieldBegins: void 0,
+		codeDepthStack: [],
+		complexFieldBaseDepth: 0
 	};
 	const preserved = dispatchChildrenWithContext({
 		element: paraElement,
@@ -1174,7 +1242,8 @@ function parseParagraph(node, styles, theme, numbering, rels = null, media = nul
 				}
 				if (implicitChildLevelAdvances > 0) listRendering.implicitChildLevelAdvances = implicitChildLevelAdvances;
 				paragraph.listRendering = listRendering;
-				const chainInd = numPrFromStyle ? styleChainInd(paragraph.formatting?.styleId, styles) : {
+				const styleInd = styleChainInd(paragraph.formatting?.styleId ?? (styles ? resolveDefaultParagraphStyle(styles.values())?.styleId : void 0), styles);
+				const chainInd = numPrFromStyle ? styleInd : {
 					left: false,
 					firstLine: false
 				};
@@ -1183,8 +1252,8 @@ function parseParagraph(node, styles, theme, numbering, rels = null, media = nul
 					const directInd = pPr ? findChild(pPr, "w", "ind") : null;
 					const hasDirectLeft = hasAttributeAnySpelling(directInd, "CT_Ind @left");
 					const hasDirectFirstLineOrHanging = parseNumericAttribute(directInd, "w", "firstLine") !== void 0 || parseNumericAttribute(directInd, "w", "hanging") !== void 0;
-					if (!hasDirectLeft && !chainInd.left && level.pPr.indentLeft !== void 0) paragraph.formatting.indentLeft = level.pPr.indentLeft;
-					if (!hasDirectFirstLineOrHanging && !chainInd.firstLine && numberingLevelHasMarkerSlot(level)) {
+					if (!hasDirectLeft && !chainInd.left && level.pPr.indentLeft !== void 0 && (level.pPr.indentLeft !== 0 || styleInd.left)) paragraph.formatting.indentLeft = level.pPr.indentLeft;
+					if (!hasDirectFirstLineOrHanging && !chainInd.firstLine && numberingLevelHasMarkerSlot(level) && (level.pPr.indentFirstLine !== 0 || styleInd.firstLine)) {
 						if (level.pPr.indentFirstLine !== void 0) paragraph.formatting.indentFirstLine = level.pPr.indentFirstLine;
 						if (level.pPr.hangingIndent !== void 0) paragraph.formatting.hangingIndent = level.pPr.hangingIndent;
 					}

package/dist/docx/paragraphPropertySource.d.ts

@@ -1,7 +1,7 @@
 import { document_d_exports } from "../types/document.js";
 import { ModeledParagraphFormattingEmission } from "../internal/paragraphFormattingSerialization.js";
-import { Transaction } from "prosemirror-state";
 import { Fragment, Mark, Node, NodeType } from "prosemirror-model";
+import { Transaction } from "prosemirror-state";
 //#region src/docx/paragraphPropertySource.d.ts
 declare const paragraphPropertySourceEmissionFingerprint: unique symbol;
 type ParagraphPropertySource = Readonly<{
@@ -97,6 +97,8 @@ type JoinProseParagraphsWithRightPropertySourceOptions = {
 };
 /** Join adjacent paragraphs when revision semantics deliberately select the right owner. */
 declare const joinProseParagraphsWithRightPropertySource: ({ attrs, pos, transaction }: JoinProseParagraphsWithRightPropertySourceOptions) => Transaction;
+/** Record the property owners deliberately selected by a batch of paragraph joins. */
+declare const markParagraphPropertySourceTransfers: (transaction: Transaction, transfers: readonly ParagraphPropertySourceTransfer[]) => void;
 declare const getExplicitParagraphPropertySourceTransfers: (transaction: Transaction) => readonly ParagraphPropertySourceTransfer[];
 type RecreateProseNodeOptions = {
   attrs?: Node["attrs"];
@@ -134,4 +136,4 @@ declare const linkParagraphPropertySourceCandidate: (target: document_d_exports.
 declare const getParagraphPropertySourceCandidate: (paragraph: document_d_exports.Paragraph) => document_d_exports.Paragraph | undefined;
 declare const getParagraphPropertySourceTransferId: (paragraph: document_d_exports.Paragraph) => string | undefined;
 //#endregion
-export { DecodedTableCellParagraphSourcePayload, PARAGRAPH_PROPERTY_SOURCE_VALIDATION_CODES, PROSE_PARAGRAPH_SOURCE_CONTRACT_ATTR, PROSE_PARAGRAPH_SOURCE_TOKEN_ATTR, ParagraphPropertySourceValidationCode, ParagraphPropertySourceValidationError, TABLE_CELL_PARAGRAPH_SOURCE_BINDING_ATTR, TABLE_CELL_PARAGRAPH_SOURCE_PAYLOAD_ERROR_CLASSIFICATIONS, TableCellParagraphPropertySourceBinding, TableCellParagraphSourcePayloadErrorClassification, assignDocumentParagraphPropertySourceContract, assignParagraphPropertySource, cloneDocumentWithParagraphPropertySources, cloneParagraphWithPropertySource, cloneParagraphWithoutPropertySource, cloneTableCellsWithParagraphPropertyCaptures, copyDocumentParagraphPropertySourceContract, copyDocumentParagraphPropertySources, copyParagraphPropertyCapture, copyParagraphPropertySource, createProseParagraphWithPropertySource, decodeTableCellParagraphSourcePayload, getDocumentParagraphPropertySourceContract, getExplicitParagraphPropertySourceTransfers, getParagraphPropertySource, getParagraphPropertySourceCandidate, getParagraphPropertySourceToken, getParagraphPropertySourceTransferId, getProseDocumentParagraphPropertySourceContract, getProseParagraphPropertySourceToken, isParagraphPropertySourceToken, joinProseParagraphsWithRightPropertySource, linkParagraphPropertySourceCandidate, linkProseParagraphPropertySource, paragraphPropertySourceBelongsToDocument, paragraphPropertySourceMatchesEmission, paragraphPropertySourceTokenMatchesContract, proseParagraphAttrsWithoutPropertySource, recreateProseNodeWithDetachedParagraphPropertySource, recreateProseNodeWithParagraphPropertySource, restoreTableCellsWithParagraphPropertySources, setProseParagraphMarkupWithPropertySource, transferProseParagraphPropertySource, transportTableCellsWithParagraphPropertySources, visitDocumentStoryParagraphs, visitTableCellParagraphPropertySourceBindings };
+export { DecodedTableCellParagraphSourcePayload, PARAGRAPH_PROPERTY_SOURCE_VALIDATION_CODES, PROSE_PARAGRAPH_SOURCE_CONTRACT_ATTR, PROSE_PARAGRAPH_SOURCE_TOKEN_ATTR, ParagraphPropertySourceTransfer, ParagraphPropertySourceValidationCode, ParagraphPropertySourceValidationError, TABLE_CELL_PARAGRAPH_SOURCE_BINDING_ATTR, TABLE_CELL_PARAGRAPH_SOURCE_PAYLOAD_ERROR_CLASSIFICATIONS, TableCellParagraphPropertySourceBinding, TableCellParagraphSourcePayloadErrorClassification, assignDocumentParagraphPropertySourceContract, assignParagraphPropertySource, cloneDocumentWithParagraphPropertySources, cloneParagraphWithPropertySource, cloneParagraphWithoutPropertySource, cloneTableCellsWithParagraphPropertyCaptures, copyDocumentParagraphPropertySourceContract, copyDocumentParagraphPropertySources, copyParagraphPropertyCapture, copyParagraphPropertySource, createProseParagraphWithPropertySource, decodeTableCellParagraphSourcePayload, getDocumentParagraphPropertySourceContract, getExplicitParagraphPropertySourceTransfers, getParagraphPropertySource, getParagraphPropertySourceCandidate, getParagraphPropertySourceToken, getParagraphPropertySourceTransferId, getProseDocumentParagraphPropertySourceContract, getProseParagraphPropertySourceToken, isParagraphPropertySourceToken, joinProseParagraphsWithRightPropertySource, linkParagraphPropertySourceCandidate, linkProseParagraphPropertySource, markParagraphPropertySourceTransfers, paragraphPropertySourceBelongsToDocument, paragraphPropertySourceMatchesEmission, paragraphPropertySourceTokenMatchesContract, proseParagraphAttrsWithoutPropertySource, recreateProseNodeWithDetachedParagraphPropertySource, recreateProseNodeWithParagraphPropertySource, restoreTableCellsWithParagraphPropertySources, setProseParagraphMarkupWithPropertySource, transferProseParagraphPropertySource, transportTableCellsWithParagraphPropertySources, visitDocumentStoryParagraphs, visitTableCellParagraphPropertySourceBindings };

package/dist/docx/paragraphPropertySource.js

@@ -261,6 +261,7 @@ const invalidTableCellParagraphSourcePayload = (classification, path) => {
 };
 const tableCellBlockTraversalByType = {
 	blockSdt: "blockSdt",
+	blockCustomXml: "blockSdt",
 	bookmarkEnd: "leaf",
 	bookmarkStart: "leaf",
 	paragraph: "paragraph",
@@ -634,6 +635,11 @@ const joinProseParagraphsWithRightPropertySource = ({ attrs, pos, transaction })
 	}]);
 	return transaction;
 };
+/** Record the property owners deliberately selected by a batch of paragraph joins. */
+const markParagraphPropertySourceTransfers = (transaction, transfers) => {
+	if (transfers.length === 0) return;
+	transaction.setMeta(paragraphPropertySourceTransfersKey, [...transaction.getMeta(paragraphPropertySourceTransfersKey) ?? [], ...transfers]);
+};
 const getExplicitParagraphPropertySourceTransfers = (transaction) => transaction.getMeta(paragraphPropertySourceTransfersKey) ?? [];
 /** Carry a parser-linked paragraph owner across an immutable PM node rebuild. */
 const copyProseParagraphPropertySource = (target, source) => {
@@ -673,7 +679,9 @@ const proseParagraphAttrsWithoutPropertySource = (attrs) => {
 /** Replace paragraph markup without losing its private parser-owner link. */
 const setProseParagraphMarkupWithPropertySource = ({ attrs, ownership, pos, transaction }) => {
 	const source = transaction.doc.nodeAt(pos);
-	transaction.setNodeMarkup(pos, void 0, attrs);
+	if (source) {
+		for (const [name, value] of Object.entries(attrs)) if (source.attrs[name] !== value) transaction.setNodeAttribute(pos, name, value);
+	} else transaction.setNodeMarkup(pos, void 0, attrs);
 	const target = transaction.doc.nodeAt(pos);
 	if (!source || !target) return;
 	if (ownership === "preserve") {
@@ -703,4 +711,4 @@ const linkParagraphPropertySourceCandidate = (target, source) => {
 const getParagraphPropertySourceCandidate = (paragraph) => paragraphPropertySourceCandidates.get(paragraph);
 const getParagraphPropertySourceTransferId = (paragraph) => paragraphPropertySourceTransferIds.get(paragraph);
 //#endregion
-export { PARAGRAPH_PROPERTY_SOURCE_VALIDATION_CODES, PROSE_PARAGRAPH_SOURCE_CONTRACT_ATTR, PROSE_PARAGRAPH_SOURCE_TOKEN_ATTR, ParagraphPropertySourceValidationError, TABLE_CELL_PARAGRAPH_SOURCE_BINDING_ATTR, TABLE_CELL_PARAGRAPH_SOURCE_PAYLOAD_ERROR_CLASSIFICATIONS, assignDocumentParagraphPropertySourceContract, assignParagraphPropertySource, cloneDocumentWithParagraphPropertySources, cloneParagraphWithPropertySource, cloneParagraphWithoutPropertySource, cloneTableCellsWithParagraphPropertyCaptures, copyDocumentParagraphPropertySourceContract, copyDocumentParagraphPropertySources, copyParagraphPropertyCapture, copyParagraphPropertySource, createProseParagraphWithPropertySource, decodeTableCellParagraphSourcePayload, getDocumentParagraphPropertySourceContract, getExplicitParagraphPropertySourceTransfers, getParagraphPropertySource, getParagraphPropertySourceCandidate, getParagraphPropertySourceToken, getParagraphPropertySourceTransferId, getProseDocumentParagraphPropertySourceContract, getProseParagraphPropertySourceToken, isParagraphPropertySourceToken, joinProseParagraphsWithRightPropertySource, linkParagraphPropertySourceCandidate, linkProseParagraphPropertySource, paragraphPropertySourceBelongsToDocument, paragraphPropertySourceMatchesEmission, paragraphPropertySourceTokenMatchesContract, proseParagraphAttrsWithoutPropertySource, recreateProseNodeWithDetachedParagraphPropertySource, recreateProseNodeWithParagraphPropertySource, restoreTableCellsWithParagraphPropertySources, setProseParagraphMarkupWithPropertySource, transferProseParagraphPropertySource, transportTableCellsWithParagraphPropertySources, visitDocumentStoryParagraphs, visitTableCellParagraphPropertySourceBindings };
+export { PARAGRAPH_PROPERTY_SOURCE_VALIDATION_CODES, PROSE_PARAGRAPH_SOURCE_CONTRACT_ATTR, PROSE_PARAGRAPH_SOURCE_TOKEN_ATTR, ParagraphPropertySourceValidationError, TABLE_CELL_PARAGRAPH_SOURCE_BINDING_ATTR, TABLE_CELL_PARAGRAPH_SOURCE_PAYLOAD_ERROR_CLASSIFICATIONS, assignDocumentParagraphPropertySourceContract, assignParagraphPropertySource, cloneDocumentWithParagraphPropertySources, cloneParagraphWithPropertySource, cloneParagraphWithoutPropertySource, cloneTableCellsWithParagraphPropertyCaptures, copyDocumentParagraphPropertySourceContract, copyDocumentParagraphPropertySources, copyParagraphPropertyCapture, copyParagraphPropertySource, createProseParagraphWithPropertySource, decodeTableCellParagraphSourcePayload, getDocumentParagraphPropertySourceContract, getExplicitParagraphPropertySourceTransfers, getParagraphPropertySource, getParagraphPropertySourceCandidate, getParagraphPropertySourceToken, getParagraphPropertySourceTransferId, getProseDocumentParagraphPropertySourceContract, getProseParagraphPropertySourceToken, isParagraphPropertySourceToken, joinProseParagraphsWithRightPropertySource, linkParagraphPropertySourceCandidate, linkProseParagraphPropertySource, markParagraphPropertySourceTransfers, paragraphPropertySourceBelongsToDocument, paragraphPropertySourceMatchesEmission, paragraphPropertySourceTokenMatchesContract, proseParagraphAttrsWithoutPropertySource, recreateProseNodeWithDetachedParagraphPropertySource, recreateProseNodeWithParagraphPropertySource, restoreTableCellsWithParagraphPropertySources, setProseParagraphMarkupWithPropertySource, transferProseParagraphPropertySource, transportTableCellsWithParagraphPropertySources, visitDocumentStoryParagraphs, visitTableCellParagraphPropertySourceBindings };

package/dist/docx/paragraphTraversal.js

@@ -126,6 +126,7 @@ const visitBlockTreeRecords = (blocks, visit) => {
 				}
 				break;
 			case "blockSdt":
+			case "blockCustomXml":
 				visitBlocks(block.content);
 				break;
 			case "preservedBlock":

package/dist/docx/parseWarningMessage.js

@@ -40,7 +40,10 @@ const PARSE_WARNING_MESSAGES = {
 	[PARSE_WARNING_CODES.outlineDashOutsideEnum]: (warning) => `Kept outline dash${quoted(warning.value)} verbatim${where(warning)}; ST_PresetLineDashVal does not declare it.`,
 	[PARSE_WARNING_CODES.styleSetDuplicateStyleId]: (warning) => `Dropped a style repeating the id${quoted(warning.value)} another style in the set already defines.`,
 	[PARSE_WARNING_CODES.styleSetInitialStyleMissing]: (warning) => `The style set names initial paragraph style${quoted(warning.value)}, which it does not contain; used the set's default instead${where(warning)}.`,
-	[PARSE_WARNING_CODES.pageBreakProjectionApproximated]: (warning) => `${warning.detail ?? "An explicit page break is laid out approximately"}${where(warning)}.`
+	[PARSE_WARNING_CODES.pageBreakProjectionApproximated]: (warning) => `${warning.detail ?? "An explicit page break is laid out approximately"}${where(warning)}.`,
+	[PARSE_WARNING_CODES.altChunkUnsupported]: (warning) => `Preserved ${plural(warning.count, "w:altChunk")}; Folio cannot import its external content. Readers show a diagnostic marker.`,
+	[PARSE_WARNING_CODES.revisionCarrierOpaque]: (warning) => `Preserved ${plural(warning.count, "block-level tracked-change wrapper")}; Folio cannot interpret its block content, so that content is unavailable to readers or edits.`,
+	[PARSE_WARNING_CODES.nestedRowOpaque]: (warning) => `Preserved ${plural(warning.count, "nested table row")} as opaque markup; readers show a diagnostic marker for its content.`
 };
 const formatParseWarning = (warning) => PARSE_WARNING_MESSAGES[warning.code](warning);
 const formatParseWarnings = (warnings) => warnings.map(formatParseWarning);

package/dist/docx/parser.js

@@ -1,6 +1,7 @@
 import { toArrayBuffer } from "../utils/docxInput.js";
 import { loadFontsWithMapping } from "../utils/fontLoader.js";
 import { MAX_PACKAGE_TIFF_PIXELS, convertTiffToPngDataUrl, isTiffMimeType } from "../utils/tiffConverter.js";
+import { ALT_CHUNK_PLAIN_TEXT_LIMIT_BYTES, getAltChunkRelationshipId, isAltChunkMarkup } from "./altChunk.js";
 import { DUPLICATE_COMMENT_ID_WARNING, normalizeCommentIds } from "./commentIdNormalization.js";
 import { parseComments } from "./commentParser.js";
 import { DANGLING_COMMENT_REFERENCE_WARNING, UNBALANCED_COMMENT_RANGE_WARNING, normalizeCommentReferences } from "./commentReferenceNormalization.js";
@@ -13,13 +14,14 @@ import { DocxEncryptionError } from "./encryption/errors.js";
 import { parseFontTable } from "./fontTableParser.js";
 import { parseEndnotes, parseFootnotes } from "./footnoteParser.js";
 import { parseFooter, parseHeader } from "./headerFooterParser.js";
-import { DANGLING_FOOTER_REFERENCE_WARNING, DANGLING_HEADER_REFERENCE_WARNING, normalizeHeaderFooterReferences } from "./headerFooterReferenceNormalization.js";
-import { assignHeaderFooterVerbatimXml, refreshHeaderFooterVerbatimFingerprint } from "./headerFooterVerbatim.js";
+import { DANGLING_FOOTER_REFERENCE_WARNING, DANGLING_HEADER_REFERENCE_WARNING, assignHeaderFooterRoles, normalizeHeaderFooterReferences } from "./headerFooterReferenceNormalization.js";
+import { assignHeaderFooterVerbatimXml, canReplayHeaderFooterVerbatim, refreshHeaderFooterVerbatimFingerprint } from "./headerFooterVerbatim.js";
 import { extractMetafileRaster, isMetafileMimeType } from "./metafileRaster.js";
 import { renderEmfSvg } from "./metafileSvg.js";
 import { DocxModelValidationError, formatDocumentModelIssues, validateFolioDocumentModel } from "./modelValidation.js";
 import { parseNumbering } from "./numberingParser.js";
 import { UNNUMBERED_PARAGRAPH_WARNING, UNNUMBERED_STYLE_WARNING, normalizeNumberingReferences, normalizeStyleNumberingReferences } from "./numberingReferenceNormalization.js";
+import { countOpaqueRevisionWrappers } from "./opaqueCarrier.js";
 import { assignDocumentParagraphPropertySourceContract } from "./paragraphPropertySource.js";
 import { createParseWarningCollector } from "./parseContext.js";
 import { formatParseWarnings } from "./parseWarningMessage.js";
@@ -30,9 +32,10 @@ import { parseSettings } from "./settingsParser.js";
 import { parseStylesPackage } from "./styleParser.js";
 import { applyThemeFontLang, parseTheme } from "./themeParser.js";
 import { UNBALANCED_MOVE_RANGE_WARNING, normalizeTrackedMoveRanges } from "./trackedMoveRangeNormalization.js";
-import { getMediaMimeType, mediaToDataUrl, unzipDocx } from "./unzip.js";
+import { getEntryUncompressedSize, getMediaMimeType, mediaToDataUrl, unzipDocx } from "./unzip.js";
+import { getAttribute, getChildElements, getLocalName, parseXmlDocument } from "./xmlParser.js";
 import { FOLIO_XML_RESOURCE_LIMITS } from "./xmlResourceLimits.js";
-import { TaggedError } from "better-result";
+import { TaggedError, panic } from "better-result";
 import { PARSE_WARNING_CODES } from "@stll/docx-core/model";
 //#region src/docx/parser.ts
 /**
@@ -63,6 +66,101 @@ const sha256Hex = async (buffer) => {
 	for (const byte of bytes) digest += byte.toString(16).padStart(2, "0");
 	return digest;
 };
+const countAltChunks = (blocks) => {
+	let count = 0;
+	for (const block of blocks) switch (block.type) {
+		case "preservedBlock":
+			count += isAltChunkMarkup(block.xml) ? 1 : 0;
+			break;
+		case "table":
+			for (const row of block.rows) for (const cell of row.cells) count += countAltChunks(cell.content);
+			break;
+		case "blockSdt":
+		case "blockCustomXml":
+			count += countAltChunks(block.content);
+			break;
+		case "paragraph":
+		case "bookmarkStart":
+		case "bookmarkEnd": break;
+		default: panic(`Unsupported block while counting altChunks: ${JSON.stringify(block)}`);
+	}
+	return count;
+};
+const ALT_CHUNK_RELATIONSHIP_TYPE = "http://schemas.openxmlformats.org/officeDocument/2006/relationships/aFChunk";
+const ALT_CHUNK_STRICT_RELATIONSHIP_TYPE = "http://purl.oclc.org/ooxml/officeDocument/relationships/aFChunk";
+const CONTENT_TYPES_NAMESPACE_URI = "http://schemas.openxmlformats.org/package/2006/content-types";
+const contentTypeOf = (contentTypesXml, partPath) => {
+	if (!contentTypesXml) return;
+	const root = parseXmlDocument(contentTypesXml);
+	if (!root || root.namespaceUri !== CONTENT_TYPES_NAMESPACE_URI) return;
+	const extension = partPath.slice(partPath.lastIndexOf(".") + 1).toLowerCase();
+	let defaultContentType;
+	for (const child of getChildElements(root)) {
+		if (child.namespaceUri !== CONTENT_TYPES_NAMESPACE_URI) continue;
+		const name = getLocalName(child.name);
+		const contentType = getAttribute(child, null, "ContentType");
+		if (name === "Override" && getAttribute(child, null, "PartName") === `/${partPath}`) return contentType ?? void 0;
+		if (name === "Default" && getAttribute(child, null, "Extension")?.toLowerCase() === extension) defaultContentType = contentType ?? void 0;
+	}
+	return defaultContentType;
+};
+const extractAltChunkPlainText = async (xml, context) => {
+	const relationshipId = getAltChunkRelationshipId(xml);
+	const relationship = relationshipId === void 0 ? void 0 : context.relationships.get(relationshipId);
+	if (!relationship || relationship.targetMode === "External" || relationship.type !== ALT_CHUNK_RELATIONSHIP_TYPE && relationship.type !== ALT_CHUNK_STRICT_RELATIONSHIP_TYPE) return;
+	const partPath = resolveRelativePath(context.relationshipsPath, relationship.target);
+	if (contentTypeOf(context.raw.contentTypesXml, partPath)?.split(";")[0]?.trim().toLowerCase() !== "text/plain") return;
+	const file = context.raw.originalZip.file(partPath);
+	if (!file) return;
+	const declaredSize = getEntryUncompressedSize(file);
+	if (declaredSize === null || declaredSize > context.budget.remainingBytes) return void 0;
+	const bytes = await file.async("uint8array");
+	if (bytes.byteLength > context.budget.remainingBytes) return;
+	context.budget.remainingBytes -= bytes.byteLength;
+	try {
+		return new TextDecoder("utf-8", { fatal: true }).decode(bytes);
+	} catch {
+		return;
+	}
+};
+const attachAltChunkPlainText = async (blocks, context) => {
+	for (const block of blocks) {
+		if (block.type === "preservedBlock" && isAltChunkMarkup(block.xml)) {
+			const readerText = await extractAltChunkPlainText(block.xml, context);
+			if (readerText !== void 0) block.readerText = readerText;
+			continue;
+		}
+		if (block.type === "table") {
+			for (const row of block.rows) for (const cell of row.cells) await attachAltChunkPlainText(cell.content, context);
+			continue;
+		}
+		if (block.type === "blockSdt" || block.type === "blockCustomXml") await attachAltChunkPlainText(block.content, context);
+	}
+};
+const countOpaqueRevisionCarriers = (blocks) => {
+	let count = 0;
+	for (const block of blocks) switch (block.type) {
+		case "preservedBlock":
+			count += countOpaqueRevisionWrappers(block.xml);
+			break;
+		case "table":
+			count += block.preserved?.children?.reduce((total, child) => total + countOpaqueRevisionWrappers(child.xml), 0) ?? 0;
+			for (const row of block.rows) {
+				count += row.preserved?.children?.reduce((total, child) => total + countOpaqueRevisionWrappers(child.xml), 0) ?? 0;
+				for (const cell of row.cells) count += countOpaqueRevisionCarriers(cell.content);
+			}
+			break;
+		case "blockSdt":
+		case "blockCustomXml":
+			count += countOpaqueRevisionCarriers(block.content);
+			break;
+		case "paragraph":
+		case "bookmarkStart":
+		case "bookmarkEnd": break;
+		default: panic(`Unsupported block while counting opaque revisions: ${JSON.stringify(block)}`);
+	}
+	return count;
+};
 /**
 * Parse a DOCX file into a complete Document model
 *
@@ -210,6 +308,11 @@ async function parseDocxWithPreviewBudget(input, options, previewBudget) {
 			code: DANGLING_FOOTER_REFERENCE_WARNING,
 			count: headerFooterReferenceNormalization.removedDanglingFooterReferences
 		});
+		assignHeaderFooterRoles({
+			documentBody,
+			...headers !== void 0 ? { headers } : {},
+			...footers !== void 0 ? { footers } : {}
+		});
 		const numberingReferenceNormalization = normalizeNumberingReferences({
 			documentBody,
 			numbering,
@@ -272,27 +375,77 @@ async function parseDocxWithPreviewBudget(input, options, previewBudget) {
 		} else onProgress("Skipping font loading", 95);
 		onProgress("Assembling document...", 95);
 		const properties = timeStage("coreProperties", () => parseCoreProperties(raw.corePropsXml));
+		const pkg = {
+			conformanceClass: detectDocxConformanceClass(raw.documentXml),
+			document: documentBody,
+			settings,
+			...styleDefinitions !== void 0 ? { styles: styleDefinitions } : {},
+			theme,
+			numbering: numbering.definitions,
+			...fontTable ? { fontTable } : {},
+			...headers !== void 0 ? { headers } : {},
+			...footers !== void 0 ? { footers } : {},
+			...footnotes !== void 0 ? { footnotes } : {},
+			...endnotes !== void 0 ? { endnotes } : {},
+			relationships: rels,
+			media,
+			...properties !== void 0 ? { properties } : {}
+		};
 		const document = {
-			package: {
-				conformanceClass: detectDocxConformanceClass(raw.documentXml),
-				document: documentBody,
-				settings,
-				...styleDefinitions !== void 0 ? { styles: styleDefinitions } : {},
-				theme,
-				numbering: numbering.definitions,
-				...fontTable ? { fontTable } : {},
-				...headers !== void 0 ? { headers } : {},
-				...footers !== void 0 ? { footers } : {},
-				...footnotes !== void 0 ? { footnotes } : {},
-				...endnotes !== void 0 ? { endnotes } : {},
-				relationships: rels,
-				media,
-				...properties !== void 0 ? { properties } : {}
-			},
+			package: pkg,
 			originalBuffer: raw.originalBuffer,
 			...templateVariables !== void 0 ? { templateVariables } : {},
 			...requiredFonts.length > 0 ? { requiredFonts } : {}
 		};
+		const altChunkBudget = { remainingBytes: ALT_CHUNK_PLAIN_TEXT_LIMIT_BYTES };
+		await attachAltChunkPlainText(pkg.document.content, {
+			relationships: rels,
+			relationshipsPath: DOCUMENT_RELATIONSHIPS_PATH,
+			raw,
+			budget: altChunkBudget
+		});
+		const altChunkContextForPart = (partPath) => {
+			const relationshipsPath = getRelationshipsPathForPart(partPath);
+			const relationshipsXml = getMapCaseInsensitive(raw.allXml, relationshipsPath);
+			return {
+				relationships: relationshipsXml ? parseRelationships(relationshipsXml) : /* @__PURE__ */ new Map(),
+				relationshipsPath,
+				raw,
+				budget: altChunkBudget
+			};
+		};
+		for (const [rId, header] of pkg.headers ?? []) {
+			const relationship = rels.get(rId);
+			if (relationship?.target) {
+				const replayable = canReplayHeaderFooterVerbatim(header);
+				await attachAltChunkPlainText(header.content, altChunkContextForPart(getRelationshipPartPath(relationship.target)));
+				if (replayable) refreshHeaderFooterVerbatimFingerprint(header);
+			}
+		}
+		for (const [rId, footer] of pkg.footers ?? []) {
+			const relationship = rels.get(rId);
+			if (relationship?.target) {
+				const replayable = canReplayHeaderFooterVerbatim(footer);
+				await attachAltChunkPlainText(footer.content, altChunkContextForPart(getRelationshipPartPath(relationship.target)));
+				if (replayable) refreshHeaderFooterVerbatimFingerprint(footer);
+			}
+		}
+		const footnoteContext = altChunkContextForPart("word/footnotes.xml");
+		for (const footnote of pkg.footnotes ?? []) await attachAltChunkPlainText(footnote.content, footnoteContext);
+		const endnoteContext = altChunkContextForPart("word/endnotes.xml");
+		for (const endnote of pkg.endnotes ?? []) await attachAltChunkPlainText(endnote.content, endnoteContext);
+		const altChunkCount = countAltChunks(pkg.document.content) + [...pkg.headers?.values() ?? []].reduce((total, part) => total + countAltChunks(part.content), 0) + [...pkg.footers?.values() ?? []].reduce((total, part) => total + countAltChunks(part.content), 0) + (pkg.footnotes ?? []).reduce((total, part) => total + countAltChunks(part.content), 0) + (pkg.endnotes ?? []).reduce((total, part) => total + countAltChunks(part.content), 0);
+		if (altChunkCount > 0) parseContext.warn({
+			code: PARSE_WARNING_CODES.altChunkUnsupported,
+			element: "w:altChunk",
+			count: altChunkCount
+		});
+		const opaqueRevisionCount = countOpaqueRevisionCarriers(pkg.document.content) + [...pkg.headers?.values() ?? []].reduce((total, part) => total + countOpaqueRevisionCarriers(part.content), 0) + [...pkg.footers?.values() ?? []].reduce((total, part) => total + countOpaqueRevisionCarriers(part.content), 0) + (pkg.footnotes ?? []).reduce((total, part) => total + countOpaqueRevisionCarriers(part.content), 0) + (pkg.endnotes ?? []).reduce((total, part) => total + countOpaqueRevisionCarriers(part.content), 0);
+		if (opaqueRevisionCount > 0) parseContext.warn({
+			code: PARSE_WARNING_CODES.revisionCarrierOpaque,
+			element: "w:ins/w:del/w:moveFrom/w:moveTo",
+			count: opaqueRevisionCount
+		});
 		assignDocumentParagraphPropertySourceContract(document, await paragraphPropertySourceDigest);
 		previews.enforce(previewBudget);
 		const validation = validateFolioDocumentModel(document);
@@ -585,7 +738,7 @@ function fullParseDocx(buffer, onProgress) {
 * Faster than full parse when you only need variables
 */
 async function getDocxVariables(buffer) {
-	const raw = await unzipDocx(buffer);
+	const raw = await unzipDocx(buffer, {}, { verifyUnreadEntries: false });
 	if (!raw.documentXml) return [];
 	return extractAllTemplateVariables(parseDocumentBody(raw.documentXml).content);
 }
@@ -593,7 +746,7 @@ async function getDocxVariables(buffer) {
 * Get document summary without full parsing
 */
 async function getDocxSummary(buffer) {
-	const raw = await unzipDocx(buffer);
+	const raw = await unzipDocx(buffer, {}, { verifyUnreadEntries: false });
 	const variables = raw.documentXml ? extractAllTemplateVariables(parseDocumentBody(raw.documentXml).content) : [];
 	return {
 		hasDocument: raw.documentXml !== null,