doc-codec 2.4.2 → 2.5.0

package/dist/read.cjs

@@ -3,22 +3,21 @@ const require_errors = require("./errors.cjs");
 const require_bytes = require("./bytes.cjs");
 const require_detect = require("./detect.cjs");
 const require_encryption = require("./encryption.cjs");
-const require_fib_fib = require("./fib/fib.cjs");
+const require_prop_fkp = require("./prop/fkp.cjs");
+const require_style_stsh = require("./style/stsh.cjs");
+const require_text_paragraphs = require("./text/paragraphs.cjs");
 const require_text_piece_table = require("./text/piece-table.cjs");
 const require_text_characters = require("./text/characters.cjs");
-const require_text_special = require("./text/special.cjs");
-const require_prop_sprm = require("./prop/sprm.cjs");
-const require_prop_fkp = require("./prop/fkp.cjs");
-const require_prop_chp = require("./prop/chp.cjs");
-const require_prop_pap = require("./prop/pap.cjs");
+const require_table_read = require("./table/read.cjs");
+const require_headers_footers = require("./headers-footers.cjs");
+const require_fib_fib = require("./fib/fib.cjs");
 const require_prop_sep = require("./prop/sep.cjs");
-const require_style_stsh = require("./style/stsh.cjs");
 const require_style_fonts = require("./style/fonts.cjs");
 const require_list_numbering = require("./list/numbering.cjs");
-const require_table_read = require("./table/read.cjs");
+const require_notes = require("./notes.cjs");
 let archive_codec = require("archive-codec");
 //#region src/read.ts
-/** Word's own default for a new document (US Letter, one-inch margins) -- what a field this reader resolves from PlcfSed/Sepx (prop/sep.ts's readSectionProperties) falls back to when the file states nothing for it, exactly as it would fall back to Word's own implementation-dependent default for that one unstated sprm. */
+/** Word's own default for a new document (US Letter, one-inch margins) -- what a field this reader resolves from PlcfSed/Sepx (prop/sep.ts's readAllSectionProperties) falls back to when the file states nothing for it, exactly as it would fall back to Word's own implementation-dependent default for that one unstated sprm. */
 const DEFAULT_PAGE_SIZE = {
 	widthPt: 612,
 	heightPt: 792
@@ -47,170 +46,67 @@ function readDocStreams(bytes, password) {
 	}
 	const fib = require_fib_fib.parseFib(wordDocument);
 	const metadata = streams.find((stream) => stream.path === require_detect.SUMMARY_INFORMATION_STREAM);
+	const data = streams.find((stream) => stream.path === require_detect.DATA_STREAM);
 	return {
 		wordDocument,
 		table,
 		fib,
-		metadata: metadata?.bytes
+		metadata: metadata?.bytes,
+		data: data?.bytes
 	};
 }
 function readDocContent(bytes, password) {
-	const { wordDocument, table, fib, metadata } = readDocStreams(bytes, password);
+	const { wordDocument, table, fib, metadata, data } = readDocStreams(bytes, password);
 	const pieceTable = require_text_piece_table.parseClx(require_bytes.slice(table, fib.fcClx, fib.lcbClx, "Clx in the Table stream"));
 	const styles = fib.lcbStshf > 0 ? require_style_stsh.parseStsh(require_bytes.slice(table, fib.fcStshf, fib.lcbStshf, "STSH in the Table stream")) : void 0;
-	const chpxTable = new require_prop_fkp.PropertyBinTable(wordDocument, require_bytes.slice(table, fib.fcPlcfBteChpx, fib.lcbPlcfBteChpx, "PlcBteChpx in the Table stream"), "PlcBteChpx");
-	const papxTable = new require_prop_fkp.PropertyBinTable(wordDocument, require_bytes.slice(table, fib.fcPlcfBtePapx, fib.lcbPlcfBtePapx, "PlcBtePapx in the Table stream"), "PlcBtePapx");
-	const fonts = fib.lcbSttbfFfn > 0 ? require_style_fonts.parseFontTable(require_bytes.slice(table, fib.fcSttbfFfn, fib.lcbSttbfFfn, "SttbfFfn in the Table stream")) : void 0;
-	const range = require_text_characters.readTextRange(wordDocument, pieceTable, 0, fib.ccpText);
-	const entries = readParagraphs(range.text, range.fcs, {
-		chpxTable,
-		papxTable,
+	const context = {
+		chpxTable: new require_prop_fkp.PropertyBinTable(wordDocument, require_bytes.slice(table, fib.fcPlcfBteChpx, fib.lcbPlcfBteChpx, "PlcBteChpx in the Table stream"), "PlcBteChpx"),
+		papxTable: new require_prop_fkp.PropertyBinTable(wordDocument, require_bytes.slice(table, fib.fcPlcfBtePapx, fib.lcbPlcfBtePapx, "PlcBtePapx in the Table stream"), "PlcBtePapx"),
 		styles,
-		fonts,
+		dataStream: data,
+		fonts: fib.lcbSttbfFfn > 0 ? require_style_fonts.parseFontTable(require_bytes.slice(table, fib.fcSttbfFfn, fib.lcbSttbfFfn, "SttbfFfn in the Table stream")) : void 0,
 		characterProperties: /* @__PURE__ */ new Map()
-	});
-	const blocks = require_table_read.assembleBlocks(entries);
+	};
+	const range = require_text_characters.readTextRange(wordDocument, pieceTable, 0, fib.ccpText);
+	const entries = require_text_paragraphs.readParagraphs(range.text, range.fcs, context);
 	const numbering = require_list_numbering.readNumberingDefinitions(table, fib);
-	const sectionProperties = require_prop_sep.readSectionProperties(wordDocument, table, fib);
+	const sectionProperties = require_prop_sep.readAllSectionProperties(wordDocument, table, fib);
+	const entriesBySection = splitIntoSections(entries, sectionProperties);
+	const { footnotes, endnotes, comments } = require_notes.readNoteBodies(wordDocument, table, pieceTable, context, fib);
+	const headerFooterStories = require_headers_footers.readHeaderFooterStories(wordDocument, table, pieceTable, context, fib, sectionProperties.length);
 	return {
 		kind: "wordprocessing",
 		metadata: metadata === void 0 ? {} : (0, archive_codec.summaryInformationToLayoutMetadata)((0, archive_codec.readSummaryInformation)(metadata)),
-		sections: [{
+		sections: sectionProperties.map((properties, index) => ({
 			pageSize: {
-				widthPt: sectionProperties.pageWidthPt ?? DEFAULT_PAGE_SIZE.widthPt,
-				heightPt: sectionProperties.pageHeightPt ?? DEFAULT_PAGE_SIZE.heightPt
+				widthPt: properties.pageWidthPt ?? DEFAULT_PAGE_SIZE.widthPt,
+				heightPt: properties.pageHeightPt ?? DEFAULT_PAGE_SIZE.heightPt
 			},
 			margins: {
-				leftPt: sectionProperties.marginLeftPt ?? DEFAULT_MARGINS.leftPt,
-				rightPt: sectionProperties.marginRightPt ?? DEFAULT_MARGINS.rightPt,
-				topPt: sectionProperties.marginTopPt ?? DEFAULT_MARGINS.topPt,
-				bottomPt: sectionProperties.marginBottomPt ?? DEFAULT_MARGINS.bottomPt
+				leftPt: properties.marginLeftPt ?? DEFAULT_MARGINS.leftPt,
+				rightPt: properties.marginRightPt ?? DEFAULT_MARGINS.rightPt,
+				topPt: properties.marginTopPt ?? DEFAULT_MARGINS.topPt,
+				bottomPt: properties.marginBottomPt ?? DEFAULT_MARGINS.bottomPt
 			},
-			blocks
-		}],
-		numbering
+			blocks: require_table_read.assembleBlocks(entriesBySection[index] ?? [])
+		})),
+		numbering,
+		footnotes,
+		endnotes,
+		comments,
+		headerFooterStories
 	};
 }
-function readParagraphs(text, fcs, context) {
-	const entries = [];
-	let start = 0;
-	for (let index = 0; index < text.length; index += 1) {
-		const code = text.charCodeAt(index);
-		if (!require_text_special.endsParagraph(code)) continue;
-		const markFc = fcs[index];
-		if (markFc === void 0) throw new require_errors.DocFormatError(`character ${index} has no byte offset, so its paragraph's properties cannot be located`);
-		entries.push(buildParagraph(text.slice(start, index), fcs.slice(start, index), markFc, code, context));
-		start = index + 1;
-	}
-	if (start < text.length) {
-		const firstFc = fcs[start];
-		if (firstFc === void 0) throw new require_errors.DocFormatError(`character ${start} has no byte offset, so the trailing paragraph's properties cannot be located`);
-		entries.push(buildParagraph(text.slice(start), fcs.slice(start), firstFc, 13, context));
-	}
-	return entries;
-}
-function buildParagraph(text, fcs, propertyFc, terminator, context) {
-	const papx = context.papxTable.papx(propertyFc);
-	const properties = {};
-	const grpprl = papx !== void 0 ? require_prop_sprm.readGrpprl(papx.grpprl) : [];
-	if (papx !== void 0) {
-		properties.istd = papx.istd;
-		if (context.styles !== void 0) require_prop_pap.applyParagraphSprms(require_style_stsh.resolveStyleFormatting(context.styles, papx.istd).paragraphPrls, properties);
-		require_prop_pap.applyParagraphSprms(grpprl, properties);
-	}
-	return {
-		paragraph: {
-			kind: "paragraph",
-			runs: buildRuns(text, fcs, context, papx?.istd),
-			...paragraphAttributes(properties, context)
-		},
-		properties,
-		grpprl,
-		terminator
-	};
-}
-function paragraphAttributes(properties, context) {
-	const attributes = {};
-	const istd = properties.istd;
-	if (istd !== void 0) {
-		const style = context.styles?.styles[istd];
-		if (style !== void 0 && style.name !== "") attributes.styleId = style.name;
-		const headingLevel = require_style_stsh.headingLevelFromIstd(istd);
-		if (headingLevel !== void 0) attributes.headingLevel = headingLevel;
-	}
-	if (attributes.headingLevel === void 0 && properties.outlineLevel !== void 0) attributes.headingLevel = properties.outlineLevel + 1;
-	if (properties.alignment !== void 0) attributes.alignment = properties.alignment;
-	if (properties.spacingBeforePt !== void 0) attributes.spacingBeforePt = properties.spacingBeforePt;
-	if (properties.spacingAfterPt !== void 0) attributes.spacingAfterPt = properties.spacingAfterPt;
-	if (properties.lineSpacing !== void 0) attributes.lineSpacing = properties.lineSpacing;
-	if (properties.indentLeftPt !== void 0) attributes.indentLeftPt = properties.indentLeftPt;
-	if (properties.indentRightPt !== void 0) attributes.indentRightPt = properties.indentRightPt;
-	if (properties.indentFirstLinePt !== void 0) attributes.indentFirstLinePt = properties.indentFirstLinePt;
-	if (properties.pageBreakBefore === true) attributes.pageBreakBefore = true;
-	if (properties.listId !== void 0) attributes.list = {
-		numId: String(properties.listId),
-		level: properties.listLevel ?? 0
-	};
-	return attributes;
-}
-function buildRuns(text, fcs, context, paragraphIstd) {
-	const runs = [];
-	let currentKey;
-	let currentText = "";
-	let currentProperties = {};
-	const paragraphStyleCharacterPrls = context.styles !== void 0 && paragraphIstd !== void 0 ? require_style_stsh.resolveStyleFormatting(context.styles, paragraphIstd).characterPrls : [];
-	const enclosingInstruction = [];
-	let inInstruction = false;
-	const flush = () => {
-		if (currentText !== "") runs.push({
-			text: currentText,
-			...currentProperties
-		});
-		currentText = "";
-	};
-	for (let index = 0; index < text.length; index += 1) {
-		const code = text.charCodeAt(index);
-		if (code === 19) {
-			flush();
-			enclosingInstruction.push(inInstruction);
-			inInstruction = true;
-			continue;
-		}
-		if (code === 20) {
-			inInstruction = false;
-			continue;
-		}
-		if (code === 21) {
-			inInstruction = enclosingInstruction.pop() ?? inInstruction;
-			continue;
-		}
-		if (inInstruction || require_text_special.isAnchorOnly(code)) continue;
-		const fc = fcs[index];
-		if (fc === void 0) throw new require_errors.DocFormatError(`character ${index} of a paragraph has no byte offset, so its formatting cannot be located`);
-		const grpprl = context.chpxTable.chpxGrpprl(fc);
-		const chpxKey = grpprl === void 0 ? "none" : `${grpprl.byteOffset}:${grpprl.byteLength}`;
-		const key = `${paragraphIstd ?? "none"}:${chpxKey}`;
-		if (key !== currentKey) {
-			flush();
-			currentKey = key;
-			let properties = context.characterProperties.get(key);
-			if (properties === void 0) {
-				properties = {};
-				require_prop_chp.applyCharacterSprms(paragraphStyleCharacterPrls, properties, context.fonts);
-				if (grpprl !== void 0) {
-					const runPrls = require_prop_sprm.readGrpprl(grpprl);
-					const characterIstd = require_prop_chp.characterIstdFromGrpprl(runPrls);
-					if (characterIstd !== void 0 && context.styles !== void 0) require_prop_chp.applyCharacterSprms(require_style_stsh.resolveStyleFormatting(context.styles, characterIstd).characterPrls, properties, context.fonts);
-					require_prop_chp.applyCharacterSprms(runPrls, properties, context.fonts);
-				}
-				context.characterProperties.set(key, properties);
-			}
-			currentProperties = properties;
-		}
-		currentText += String.fromCharCode(code === 11 ? 10 : code);
+function splitIntoSections(entries, sections) {
+	const groups = sections.map(() => []);
+	let sectionIndex = 0;
+	for (const entry of entries) {
+		const group = groups[sectionIndex];
+		if (group === void 0) throw new require_errors.DocFormatError(`internal defect: section index ${sectionIndex} has no group despite ${sections.length} sections`);
+		group.push(entry);
+		if (entry.endCp === sections[sectionIndex + 1]?.startCp) sectionIndex += 1;
 	}
-	flush();
-	return runs;
+	return groups;
 }
 //#endregion
 exports.readDocContent = readDocContent;

package/dist/read.d.cts

@@ -1,8 +1,8 @@
-import { t as Fib } from "./fib-DWFta19r.cjs";
-import { t as Prl } from "./sprm-DEyvdOo0.cjs";
-import { t as ParagraphProperties } from "./pap-BT1p0uzu.cjs";
-import { r as NumberingDefinitions } from "./numbering-DSpm9w6y.cjs";
-import { ContentDocument, ContentParagraph } from "document-schema.js";
+import { t as Fib } from "./fib-6BFIx153.cjs";
+import { HeaderFooterStories } from "./headers-footers.cjs";
+import { NumberingDefinitions } from "./list/numbering.cjs";
+import { NoteBodies } from "./notes.cjs";
+import { ContentDocument } from "document-schema.js";
 //#region src/read.d.ts
 interface DocStreams {
   readonly wordDocument: Uint8Array;
@@ -10,20 +10,18 @@ interface DocStreams {
   readonly fib: Fib;
   /** The raw "\x05SummaryInformation" stream bytes, or undefined when the container carries none -- a valid, spec-conformant Word Binary File need not carry document properties at all. */
   readonly metadata: Uint8Array<ArrayBuffer> | undefined;
+  /** The raw "Data" stream bytes, or undefined when the container carries none -- a valid Word Binary File with no pictures need not have one. sprmCPicLocation's operand addresses this stream (pictures.ts). */
+  readonly data: Uint8Array<ArrayBuffer> | undefined;
 }
 declare function readDocStreams(bytes: Uint8Array<ArrayBuffer>, password?: string): DocStreams;
-/** readDocContent's own return type: a ContentDocument (kind 'wordprocessing') plus numbering -- the list-level formatting (glyph/format, level-text template, start-at value) PlfLst/PlfLfo carry, which ContentListMembership has nowhere to hold. Mirrors ooxml.js's own DocxDocument.numbering exactly in field name and NumberingDefinitions' own shape (see list/numbering.ts's top comment for why it sits outside the shared schema rather than inside ContentListMembership); unlike DocxDocument, DocContent stays a genuine ContentDocument subtype (an intersection, not a fresh shape) since readDocContent already had one return type to widen rather than two to reconcile. */
+/** readDocContent's own return type: a ContentDocument (kind 'wordprocessing') plus numbering, footnotes, endnotes, comments, and headerFooterStories -- constructs [MS-DOC] carries outside the main document's own text and document-schema.js's ContentDocument has nowhere to hold. Mirrors ooxml.js's own DocxDocument in field name and shape wherever the two formats' own constructs genuinely agree (numbering/NumberingDefinitions, footnotes/endnotes/comments as plain-text `Footnote`/`Comment`); headerFooterStories is doc-codec's own shape rather than DocxDocument's path-addressed HeaderFooterPart, since [MS-DOC] has no named parts of its own for a header or footer to be identified by, only a (section, slot) position (see headers-footers.ts's own top comment). Unlike DocxDocument, DocContent stays a genuine ContentDocument subtype (an intersection, not a fresh shape) since readDocContent already had one return type to widen rather than several to reconcile. */
 type DocContent = ContentDocument & {
   readonly numbering: NumberingDefinitions;
+  readonly footnotes: NoteBodies["footnotes"];
+  readonly endnotes: NoteBodies["endnotes"];
+  readonly comments: NoteBodies["comments"];
+  readonly headerFooterStories: HeaderFooterStories;
 };
 declare function readDocContent(bytes: Uint8Array<ArrayBuffer>, password?: string): DocContent;
-/** One paragraph/cell/row-ending mark, still flat -- table/read.ts's assembleBlocks is what folds a run of these into a real ContentTable. `properties` and `grpprl` are carried alongside the already-built `paragraph` because table grouping needs sprmPFInTable/sprmPFTtp/sprmPItap (properties) and, on a row's own mark, its table-defining sgc-5 sprms (grpprl) -- neither of which survives onto a plain ContentParagraph. */
-interface ParagraphEntry {
-  readonly paragraph: ContentParagraph;
-  readonly properties: ParagraphProperties;
-  readonly grpprl: readonly Prl[];
-  /** The character that terminated this paragraph in the text stream: PARAGRAPH_MARK, CELL_MARK, or SECTION_MARK. */
-  readonly terminator: number;
-}
 //#endregion
-export { DocContent, DocStreams, ParagraphEntry, readDocContent, readDocStreams };
+export { DocContent, DocStreams, readDocContent, readDocStreams };

package/dist/read.d.ts

@@ -1,8 +1,8 @@
-import { t as Fib } from "./fib-DWFta19r.js";
-import { t as Prl } from "./sprm-DEyvdOo0.js";
-import { t as ParagraphProperties } from "./pap-DOFLMWcG.js";
-import { r as NumberingDefinitions } from "./numbering-DayjnHXY.js";
-import { ContentDocument, ContentParagraph } from "document-schema.js";
+import { t as Fib } from "./fib-6BFIx153.js";
+import { HeaderFooterStories } from "./headers-footers.js";
+import { NumberingDefinitions } from "./list/numbering.js";
+import { NoteBodies } from "./notes.js";
+import { ContentDocument } from "document-schema.js";
 //#region src/read.d.ts
 interface DocStreams {
   readonly wordDocument: Uint8Array;
@@ -10,20 +10,18 @@ interface DocStreams {
   readonly fib: Fib;
   /** The raw "\x05SummaryInformation" stream bytes, or undefined when the container carries none -- a valid, spec-conformant Word Binary File need not carry document properties at all. */
   readonly metadata: Uint8Array<ArrayBuffer> | undefined;
+  /** The raw "Data" stream bytes, or undefined when the container carries none -- a valid Word Binary File with no pictures need not have one. sprmCPicLocation's operand addresses this stream (pictures.ts). */
+  readonly data: Uint8Array<ArrayBuffer> | undefined;
 }
 declare function readDocStreams(bytes: Uint8Array<ArrayBuffer>, password?: string): DocStreams;
-/** readDocContent's own return type: a ContentDocument (kind 'wordprocessing') plus numbering -- the list-level formatting (glyph/format, level-text template, start-at value) PlfLst/PlfLfo carry, which ContentListMembership has nowhere to hold. Mirrors ooxml.js's own DocxDocument.numbering exactly in field name and NumberingDefinitions' own shape (see list/numbering.ts's top comment for why it sits outside the shared schema rather than inside ContentListMembership); unlike DocxDocument, DocContent stays a genuine ContentDocument subtype (an intersection, not a fresh shape) since readDocContent already had one return type to widen rather than two to reconcile. */
+/** readDocContent's own return type: a ContentDocument (kind 'wordprocessing') plus numbering, footnotes, endnotes, comments, and headerFooterStories -- constructs [MS-DOC] carries outside the main document's own text and document-schema.js's ContentDocument has nowhere to hold. Mirrors ooxml.js's own DocxDocument in field name and shape wherever the two formats' own constructs genuinely agree (numbering/NumberingDefinitions, footnotes/endnotes/comments as plain-text `Footnote`/`Comment`); headerFooterStories is doc-codec's own shape rather than DocxDocument's path-addressed HeaderFooterPart, since [MS-DOC] has no named parts of its own for a header or footer to be identified by, only a (section, slot) position (see headers-footers.ts's own top comment). Unlike DocxDocument, DocContent stays a genuine ContentDocument subtype (an intersection, not a fresh shape) since readDocContent already had one return type to widen rather than several to reconcile. */
 type DocContent = ContentDocument & {
   readonly numbering: NumberingDefinitions;
+  readonly footnotes: NoteBodies["footnotes"];
+  readonly endnotes: NoteBodies["endnotes"];
+  readonly comments: NoteBodies["comments"];
+  readonly headerFooterStories: HeaderFooterStories;
 };
 declare function readDocContent(bytes: Uint8Array<ArrayBuffer>, password?: string): DocContent;
-/** One paragraph/cell/row-ending mark, still flat -- table/read.ts's assembleBlocks is what folds a run of these into a real ContentTable. `properties` and `grpprl` are carried alongside the already-built `paragraph` because table grouping needs sprmPFInTable/sprmPFTtp/sprmPItap (properties) and, on a row's own mark, its table-defining sgc-5 sprms (grpprl) -- neither of which survives onto a plain ContentParagraph. */
-interface ParagraphEntry {
-  readonly paragraph: ContentParagraph;
-  readonly properties: ParagraphProperties;
-  readonly grpprl: readonly Prl[];
-  /** The character that terminated this paragraph in the text stream: PARAGRAPH_MARK, CELL_MARK, or SECTION_MARK. */
-  readonly terminator: number;
-}
 //#endregion
-export { DocContent, DocStreams, ParagraphEntry, readDocContent, readDocStreams };
+export { DocContent, DocStreams, readDocContent, readDocStreams };

package/dist/read.js

@@ -1,23 +1,22 @@
 import { DocFormatError, DocUnsupportedError } from "./errors.js";
 import { slice } from "./bytes.js";
-import { SUMMARY_INFORMATION_STREAM, WORD_DOCUMENT_STREAM } from "./detect.js";
+import { DATA_STREAM, SUMMARY_INFORMATION_STREAM, WORD_DOCUMENT_STREAM } from "./detect.js";
 import { decryptDocStreams } from "./encryption.js";
-import { parseFib, peekFibBaseFlags } from "./fib/fib.js";
+import { PropertyBinTable } from "./prop/fkp.js";
+import { parseStsh } from "./style/stsh.js";
+import { readParagraphs } from "./text/paragraphs.js";
 import { parseClx } from "./text/piece-table.js";
 import { readTextRange } from "./text/characters.js";
-import { endsParagraph, isAnchorOnly } from "./text/special.js";
-import { readGrpprl } from "./prop/sprm.js";
-import { PropertyBinTable } from "./prop/fkp.js";
-import { applyCharacterSprms, characterIstdFromGrpprl } from "./prop/chp.js";
-import { applyParagraphSprms } from "./prop/pap.js";
-import { readSectionProperties } from "./prop/sep.js";
-import { headingLevelFromIstd, parseStsh, resolveStyleFormatting } from "./style/stsh.js";
+import { assembleBlocks } from "./table/read.js";
+import { readHeaderFooterStories } from "./headers-footers.js";
+import { parseFib, peekFibBaseFlags } from "./fib/fib.js";
+import { readAllSectionProperties } from "./prop/sep.js";
 import { parseFontTable } from "./style/fonts.js";
 import { readNumberingDefinitions } from "./list/numbering.js";
-import { assembleBlocks } from "./table/read.js";
+import { readNoteBodies } from "./notes.js";
 import { readCompoundFile, readSummaryInformation, summaryInformationToLayoutMetadata } from "archive-codec";
 //#region src/read.ts
-/** Word's own default for a new document (US Letter, one-inch margins) -- what a field this reader resolves from PlcfSed/Sepx (prop/sep.ts's readSectionProperties) falls back to when the file states nothing for it, exactly as it would fall back to Word's own implementation-dependent default for that one unstated sprm. */
+/** Word's own default for a new document (US Letter, one-inch margins) -- what a field this reader resolves from PlcfSed/Sepx (prop/sep.ts's readAllSectionProperties) falls back to when the file states nothing for it, exactly as it would fall back to Word's own implementation-dependent default for that one unstated sprm. */
 const DEFAULT_PAGE_SIZE = {
 	widthPt: 612,
 	heightPt: 792
@@ -46,170 +45,67 @@ function readDocStreams(bytes, password) {
 	}
 	const fib = parseFib(wordDocument);
 	const metadata = streams.find((stream) => stream.path === SUMMARY_INFORMATION_STREAM);
+	const data = streams.find((stream) => stream.path === DATA_STREAM);
 	return {
 		wordDocument,
 		table,
 		fib,
-		metadata: metadata?.bytes
+		metadata: metadata?.bytes,
+		data: data?.bytes
 	};
 }
 function readDocContent(bytes, password) {
-	const { wordDocument, table, fib, metadata } = readDocStreams(bytes, password);
+	const { wordDocument, table, fib, metadata, data } = readDocStreams(bytes, password);
 	const pieceTable = parseClx(slice(table, fib.fcClx, fib.lcbClx, "Clx in the Table stream"));
 	const styles = fib.lcbStshf > 0 ? parseStsh(slice(table, fib.fcStshf, fib.lcbStshf, "STSH in the Table stream")) : void 0;
-	const chpxTable = new PropertyBinTable(wordDocument, slice(table, fib.fcPlcfBteChpx, fib.lcbPlcfBteChpx, "PlcBteChpx in the Table stream"), "PlcBteChpx");
-	const papxTable = new PropertyBinTable(wordDocument, slice(table, fib.fcPlcfBtePapx, fib.lcbPlcfBtePapx, "PlcBtePapx in the Table stream"), "PlcBtePapx");
-	const fonts = fib.lcbSttbfFfn > 0 ? parseFontTable(slice(table, fib.fcSttbfFfn, fib.lcbSttbfFfn, "SttbfFfn in the Table stream")) : void 0;
-	const range = readTextRange(wordDocument, pieceTable, 0, fib.ccpText);
-	const entries = readParagraphs(range.text, range.fcs, {
-		chpxTable,
-		papxTable,
+	const context = {
+		chpxTable: new PropertyBinTable(wordDocument, slice(table, fib.fcPlcfBteChpx, fib.lcbPlcfBteChpx, "PlcBteChpx in the Table stream"), "PlcBteChpx"),
+		papxTable: new PropertyBinTable(wordDocument, slice(table, fib.fcPlcfBtePapx, fib.lcbPlcfBtePapx, "PlcBtePapx in the Table stream"), "PlcBtePapx"),
 		styles,
-		fonts,
+		dataStream: data,
+		fonts: fib.lcbSttbfFfn > 0 ? parseFontTable(slice(table, fib.fcSttbfFfn, fib.lcbSttbfFfn, "SttbfFfn in the Table stream")) : void 0,
 		characterProperties: /* @__PURE__ */ new Map()
-	});
-	const blocks = assembleBlocks(entries);
+	};
+	const range = readTextRange(wordDocument, pieceTable, 0, fib.ccpText);
+	const entries = readParagraphs(range.text, range.fcs, context);
 	const numbering = readNumberingDefinitions(table, fib);
-	const sectionProperties = readSectionProperties(wordDocument, table, fib);
+	const sectionProperties = readAllSectionProperties(wordDocument, table, fib);
+	const entriesBySection = splitIntoSections(entries, sectionProperties);
+	const { footnotes, endnotes, comments } = readNoteBodies(wordDocument, table, pieceTable, context, fib);
+	const headerFooterStories = readHeaderFooterStories(wordDocument, table, pieceTable, context, fib, sectionProperties.length);
 	return {
 		kind: "wordprocessing",
 		metadata: metadata === void 0 ? {} : summaryInformationToLayoutMetadata(readSummaryInformation(metadata)),
-		sections: [{
+		sections: sectionProperties.map((properties, index) => ({
 			pageSize: {
-				widthPt: sectionProperties.pageWidthPt ?? DEFAULT_PAGE_SIZE.widthPt,
-				heightPt: sectionProperties.pageHeightPt ?? DEFAULT_PAGE_SIZE.heightPt
+				widthPt: properties.pageWidthPt ?? DEFAULT_PAGE_SIZE.widthPt,
+				heightPt: properties.pageHeightPt ?? DEFAULT_PAGE_SIZE.heightPt
 			},
 			margins: {
-				leftPt: sectionProperties.marginLeftPt ?? DEFAULT_MARGINS.leftPt,
-				rightPt: sectionProperties.marginRightPt ?? DEFAULT_MARGINS.rightPt,
-				topPt: sectionProperties.marginTopPt ?? DEFAULT_MARGINS.topPt,
-				bottomPt: sectionProperties.marginBottomPt ?? DEFAULT_MARGINS.bottomPt
+				leftPt: properties.marginLeftPt ?? DEFAULT_MARGINS.leftPt,
+				rightPt: properties.marginRightPt ?? DEFAULT_MARGINS.rightPt,
+				topPt: properties.marginTopPt ?? DEFAULT_MARGINS.topPt,
+				bottomPt: properties.marginBottomPt ?? DEFAULT_MARGINS.bottomPt
 			},
-			blocks
-		}],
-		numbering
-	};
-}
-function readParagraphs(text, fcs, context) {
-	const entries = [];
-	let start = 0;
-	for (let index = 0; index < text.length; index += 1) {
-		const code = text.charCodeAt(index);
-		if (!endsParagraph(code)) continue;
-		const markFc = fcs[index];
-		if (markFc === void 0) throw new DocFormatError(`character ${index} has no byte offset, so its paragraph's properties cannot be located`);
-		entries.push(buildParagraph(text.slice(start, index), fcs.slice(start, index), markFc, code, context));
-		start = index + 1;
-	}
-	if (start < text.length) {
-		const firstFc = fcs[start];
-		if (firstFc === void 0) throw new DocFormatError(`character ${start} has no byte offset, so the trailing paragraph's properties cannot be located`);
-		entries.push(buildParagraph(text.slice(start), fcs.slice(start), firstFc, 13, context));
-	}
-	return entries;
-}
-function buildParagraph(text, fcs, propertyFc, terminator, context) {
-	const papx = context.papxTable.papx(propertyFc);
-	const properties = {};
-	const grpprl = papx !== void 0 ? readGrpprl(papx.grpprl) : [];
-	if (papx !== void 0) {
-		properties.istd = papx.istd;
-		if (context.styles !== void 0) applyParagraphSprms(resolveStyleFormatting(context.styles, papx.istd).paragraphPrls, properties);
-		applyParagraphSprms(grpprl, properties);
-	}
-	return {
-		paragraph: {
-			kind: "paragraph",
-			runs: buildRuns(text, fcs, context, papx?.istd),
-			...paragraphAttributes(properties, context)
-		},
-		properties,
-		grpprl,
-		terminator
+			blocks: assembleBlocks(entriesBySection[index] ?? [])
+		})),
+		numbering,
+		footnotes,
+		endnotes,
+		comments,
+		headerFooterStories
 	};
 }
-function paragraphAttributes(properties, context) {
-	const attributes = {};
-	const istd = properties.istd;
-	if (istd !== void 0) {
-		const style = context.styles?.styles[istd];
-		if (style !== void 0 && style.name !== "") attributes.styleId = style.name;
-		const headingLevel = headingLevelFromIstd(istd);
-		if (headingLevel !== void 0) attributes.headingLevel = headingLevel;
-	}
-	if (attributes.headingLevel === void 0 && properties.outlineLevel !== void 0) attributes.headingLevel = properties.outlineLevel + 1;
-	if (properties.alignment !== void 0) attributes.alignment = properties.alignment;
-	if (properties.spacingBeforePt !== void 0) attributes.spacingBeforePt = properties.spacingBeforePt;
-	if (properties.spacingAfterPt !== void 0) attributes.spacingAfterPt = properties.spacingAfterPt;
-	if (properties.lineSpacing !== void 0) attributes.lineSpacing = properties.lineSpacing;
-	if (properties.indentLeftPt !== void 0) attributes.indentLeftPt = properties.indentLeftPt;
-	if (properties.indentRightPt !== void 0) attributes.indentRightPt = properties.indentRightPt;
-	if (properties.indentFirstLinePt !== void 0) attributes.indentFirstLinePt = properties.indentFirstLinePt;
-	if (properties.pageBreakBefore === true) attributes.pageBreakBefore = true;
-	if (properties.listId !== void 0) attributes.list = {
-		numId: String(properties.listId),
-		level: properties.listLevel ?? 0
-	};
-	return attributes;
-}
-function buildRuns(text, fcs, context, paragraphIstd) {
-	const runs = [];
-	let currentKey;
-	let currentText = "";
-	let currentProperties = {};
-	const paragraphStyleCharacterPrls = context.styles !== void 0 && paragraphIstd !== void 0 ? resolveStyleFormatting(context.styles, paragraphIstd).characterPrls : [];
-	const enclosingInstruction = [];
-	let inInstruction = false;
-	const flush = () => {
-		if (currentText !== "") runs.push({
-			text: currentText,
-			...currentProperties
-		});
-		currentText = "";
-	};
-	for (let index = 0; index < text.length; index += 1) {
-		const code = text.charCodeAt(index);
-		if (code === 19) {
-			flush();
-			enclosingInstruction.push(inInstruction);
-			inInstruction = true;
-			continue;
-		}
-		if (code === 20) {
-			inInstruction = false;
-			continue;
-		}
-		if (code === 21) {
-			inInstruction = enclosingInstruction.pop() ?? inInstruction;
-			continue;
-		}
-		if (inInstruction || isAnchorOnly(code)) continue;
-		const fc = fcs[index];
-		if (fc === void 0) throw new DocFormatError(`character ${index} of a paragraph has no byte offset, so its formatting cannot be located`);
-		const grpprl = context.chpxTable.chpxGrpprl(fc);
-		const chpxKey = grpprl === void 0 ? "none" : `${grpprl.byteOffset}:${grpprl.byteLength}`;
-		const key = `${paragraphIstd ?? "none"}:${chpxKey}`;
-		if (key !== currentKey) {
-			flush();
-			currentKey = key;
-			let properties = context.characterProperties.get(key);
-			if (properties === void 0) {
-				properties = {};
-				applyCharacterSprms(paragraphStyleCharacterPrls, properties, context.fonts);
-				if (grpprl !== void 0) {
-					const runPrls = readGrpprl(grpprl);
-					const characterIstd = characterIstdFromGrpprl(runPrls);
-					if (characterIstd !== void 0 && context.styles !== void 0) applyCharacterSprms(resolveStyleFormatting(context.styles, characterIstd).characterPrls, properties, context.fonts);
-					applyCharacterSprms(runPrls, properties, context.fonts);
-				}
-				context.characterProperties.set(key, properties);
-			}
-			currentProperties = properties;
-		}
-		currentText += String.fromCharCode(code === 11 ? 10 : code);
+function splitIntoSections(entries, sections) {
+	const groups = sections.map(() => []);
+	let sectionIndex = 0;
+	for (const entry of entries) {
+		const group = groups[sectionIndex];
+		if (group === void 0) throw new DocFormatError(`internal defect: section index ${sectionIndex} has no group despite ${sections.length} sections`);
+		group.push(entry);
+		if (entry.endCp === sections[sectionIndex + 1]?.startCp) sectionIndex += 1;
 	}
-	flush();
-	return runs;
+	return groups;
 }
 //#endregion
 export { readDocContent, readDocStreams };

package/dist/stsh-Bj0YSEOp.d.ts

@@ -0,0 +1,40 @@
+import { t as Prl } from "./sprm-DEyvdOo0.js";
+//#region src/style/stsh.d.ts
+/** StdfBase.sti's "0x0FFE for user-defined styles"; anything else is an application-defined style whose sti identifies it. */
+declare const STI_USER_DEFINED = 4094;
+/** StdfBase.stk, [MS-DOC] 2.9.269 -- which of ECMA-376's ST_StyleType values this style is. */
+declare const STK: {
+  readonly paragraph: 1;
+  readonly character: 2;
+  readonly table: 3;
+  readonly numbering: 4;
+};
+interface Style {
+  /** The style's index in STSH.rglpstd, which is what sprmPIstd and sprmCIstd name. */
+  readonly istd: number;
+  /** The invariant application-defined style identifier, or STI_USER_DEFINED for a style the document itself defines. */
+  readonly sti: number;
+  readonly stk: number;
+  /** The istd this style inherits from, or undefined when StdfBase.istdBase is 0x0FFF ("this style does not inherit from any other style"). */
+  readonly istdBase: number | undefined;
+  /** The style's primary name, from the STD's own Xstz. */
+  readonly name: string;
+  /** The style's own paragraph formatting -- StkParaGRLPUPX's lpUpxPapx.PAPX.grpprlPapx -- present only for a paragraph style (STK.paragraph). Undefined for every other stk: table and numbering styles carry their own formatting shapes this package does not read (see the README's own scope note). */
+  readonly grpprlPapx: readonly Prl[] | undefined;
+  /** The style's own character formatting -- a paragraph style's StkParaGRLPUPX.lpUpxChpx.CHPX.grpprlChpx (the run-level defaults a heading's own text falls back to) or a character style's StkCharGRLPUPX.lpUpxChpx.CHPX.grpprlChpx -- present for STK.paragraph and STK.character, undefined otherwise. */
+  readonly grpprlChpx: readonly Prl[] | undefined;
+}
+interface StyleSheet {
+  /** Indexed by istd; a hole is a slot [MS-DOC] permits to be empty ("A style definition can be empty, in which case cbStd MUST be 0"). */
+  readonly styles: readonly (Style | undefined)[];
+}
+declare function parseStsh(stsh: Uint8Array): StyleSheet;
+interface ResolvedStyleFormatting {
+  readonly paragraphPrls: readonly Prl[];
+  readonly characterPrls: readonly Prl[];
+}
+declare function resolveStyleFormatting(styleSheet: StyleSheet, istd: number): ResolvedStyleFormatting;
+declare function headingLevelFromIstd(istd: number): number | undefined;
+declare function buildStshForStyles(names: ReadonlyMap<number, string>): Uint8Array;
+//#endregion
+export { StyleSheet as a, parseStsh as c, Style as i, resolveStyleFormatting as l, STI_USER_DEFINED as n, buildStshForStyles as o, STK as r, headingLevelFromIstd as s, ResolvedStyleFormatting as t };