@cj-tech-master/excelts 3.1.0-canary.20260102071304.d80b29d → 4.0.0-canary.20260103013222.0e5d37f

package/dist/browser/modules/archive/parse.base.js

@@ -0,0 +1,610 @@
+import { parseDosDateTimeUTC, resolveZipLastModifiedDateFromUnixSeconds } from "./utils/timestamps.js";
+import { Duplex, PassThrough, Transform, concatUint8Arrays, pipeline } from "../stream/index.browser.js";
+import { parseTyped as parseBuffer } from "./utils/parse-buffer.js";
+import { ByteQueue } from "./byte-queue.js";
+import { indexOfUint8ArrayPattern } from "./utils/bytes.js";
+import { readUint32LE, writeUint32LE } from "./utils/binary.js";
+import { parseZipExtraFields } from "./utils/zip-extra-fields.js";
+import { CENTRAL_DIR_HEADER_SIG, END_OF_CENTRAL_DIR_SIG, LOCAL_FILE_HEADER_SIG, ZIP64_END_OF_CENTRAL_DIR_LOCATOR_SIG, ZIP64_END_OF_CENTRAL_DIR_SIG } from "./zip-constants.js";
+// Shared parseBuffer() formats
+export const CRX_HEADER_FORMAT = [
+    ["version", 4],
+    ["pubKeyLength", 4],
+    ["signatureLength", 4]
+];
+export const LOCAL_FILE_HEADER_FORMAT = [
+    ["versionsNeededToExtract", 2],
+    ["flags", 2],
+    ["compressionMethod", 2],
+    ["lastModifiedTime", 2],
+    ["lastModifiedDate", 2],
+    ["crc32", 4],
+    ["compressedSize", 4],
+    ["uncompressedSize", 4],
+    ["fileNameLength", 2],
+    ["extraFieldLength", 2]
+];
+export const DATA_DESCRIPTOR_FORMAT = [
+    ["dataDescriptorSignature", 4],
+    ["crc32", 4],
+    ["compressedSize", 4],
+    ["uncompressedSize", 4]
+];
+export const CENTRAL_DIRECTORY_FILE_HEADER_FORMAT = [
+    ["versionMadeBy", 2],
+    ["versionsNeededToExtract", 2],
+    ["flags", 2],
+    ["compressionMethod", 2],
+    ["lastModifiedTime", 2],
+    ["lastModifiedDate", 2],
+    ["crc32", 4],
+    ["compressedSize", 4],
+    ["uncompressedSize", 4],
+    ["fileNameLength", 2],
+    ["extraFieldLength", 2],
+    ["fileCommentLength", 2],
+    ["diskNumber", 2],
+    ["internalFileAttributes", 2],
+    ["externalFileAttributes", 4],
+    ["offsetToLocalFileHeader", 4]
+];
+export const END_OF_CENTRAL_DIRECTORY_FORMAT = [
+    ["diskNumber", 2],
+    ["diskStart", 2],
+    ["numberOfRecordsOnDisk", 2],
+    ["numberOfRecords", 2],
+    ["sizeOfCentralDirectory", 4],
+    ["offsetToStartOfCentralDirectory", 4],
+    ["commentLength", 2]
+];
+const textDecoder = new TextDecoder();
+const textEncoder = new TextEncoder();
+export function decodeZipEntryPath(pathBuffer) {
+    return textDecoder.decode(pathBuffer);
+}
+export function isZipUnicodeFlag(flags) {
+    return ((flags || 0) & 0x800) !== 0;
+}
+export function isZipDirectoryPath(path) {
+    if (path.length === 0) {
+        return false;
+    }
+    const last = path.charCodeAt(path.length - 1);
+    return last === 47 || last === 92;
+}
+export function getZipEntryType(path, uncompressedSize) {
+    return uncompressedSize === 0 && isZipDirectoryPath(path) ? "Directory" : "File";
+}
+export function buildZipEntryProps(path, pathBuffer, flags) {
+    return {
+        path,
+        pathBuffer,
+        flags: {
+            isUnicode: isZipUnicodeFlag(flags)
+        }
+    };
+}
+export function resolveZipEntryLastModifiedDateTime(vars, extraFields) {
+    const dosDate = vars.lastModifiedDate || 0;
+    const dosTime = vars.lastModifiedTime || 0;
+    const dosDateTime = parseDosDateTimeUTC(dosDate, dosTime);
+    const unixSecondsMtime = extraFields.mtimeUnixSeconds;
+    if (unixSecondsMtime === undefined) {
+        return dosDateTime;
+    }
+    return resolveZipLastModifiedDateFromUnixSeconds(dosDate, dosTime, unixSecondsMtime);
+}
+export const parseExtraField = parseZipExtraFields;
+export function hasDataDescriptorFlag(flags) {
+    return ((flags || 0) & 0x08) !== 0;
+}
+export function isFileSizeKnown(flags, compressedSize) {
+    return !hasDataDescriptorFlag(flags) || (compressedSize || 0) > 0;
+}
+export function autodrain(stream) {
+    const draining = stream.pipe(new Transform({
+        transform(_chunk, _encoding, callback) {
+            callback();
+        }
+    }));
+    draining.promise = () => new Promise((resolve, reject) => {
+        draining.on("finish", resolve);
+        draining.on("error", reject);
+    });
+    return draining;
+}
+/**
+ * Collects all data from a readable stream into a single Uint8Array.
+ */
+export function bufferStream(entry) {
+    return new Promise((resolve, reject) => {
+        const chunks = [];
+        const stream = new Transform({
+            transform(d, _encoding, callback) {
+                chunks.push(d);
+                callback();
+            }
+        });
+        stream.on("finish", () => {
+            resolve(chunks.length === 1 ? chunks[0] : concatUint8Arrays(chunks));
+        });
+        stream.on("error", reject);
+        entry.on("error", reject).pipe(stream);
+    });
+}
+const STR_FUNCTION = "function";
+export class PullStream extends Duplex {
+    get buffer() {
+        return this._queue.view();
+    }
+    set buffer(value) {
+        this._queue.reset(value);
+    }
+    constructor() {
+        super({ decodeStrings: false, objectMode: true });
+        this._queue = new ByteQueue();
+        this.finished = false;
+        this.on("finish", () => {
+            this.finished = true;
+            this.emit("chunk", false);
+        });
+    }
+    _write(chunk, _encoding, callback) {
+        const data = typeof chunk === "string" ? textEncoder.encode(chunk) : chunk;
+        this._queue.append(data);
+        this.cb = callback;
+        this.emit("chunk");
+    }
+    _read() { }
+    _maybeReleaseWriteCallback() {
+        if (typeof this.cb === STR_FUNCTION) {
+            const callback = this.cb;
+            this.cb = undefined;
+            callback();
+        }
+    }
+    /**
+     * The `eof` parameter is interpreted as `file_length` if the type is number
+     * otherwise (i.e. Uint8Array) it is interpreted as a pattern signaling end of stream
+     */
+    stream(eof, includeEof) {
+        const p = new PassThrough();
+        let done = false;
+        const cb = () => {
+            this._maybeReleaseWriteCallback();
+        };
+        const pull = () => {
+            let packet;
+            const available = this._queue.length;
+            if (available) {
+                if (typeof eof === "number") {
+                    const toRead = Math.min(eof, available);
+                    if (toRead > 0) {
+                        packet = this._queue.read(toRead);
+                        eof -= toRead;
+                    }
+                    done = done || eof === 0;
+                }
+                else {
+                    const view = this._queue.view();
+                    let match = indexOfUint8ArrayPattern(view, eof);
+                    if (match !== -1) {
+                        // store signature match byte offset to allow us to reference
+                        // this for zip64 offset
+                        this.match = match;
+                        if (includeEof) {
+                            match = match + eof.length;
+                        }
+                        if (match > 0) {
+                            packet = this._queue.read(match);
+                        }
+                        done = true;
+                    }
+                    else {
+                        const len = view.length - eof.length;
+                        if (len <= 0) {
+                            cb();
+                        }
+                        else {
+                            packet = this._queue.read(len);
+                        }
+                    }
+                }
+                if (packet) {
+                    p.write(packet, () => {
+                        if (this._queue.length === 0 ||
+                            (typeof eof !== "number" && eof.length && this._queue.length <= eof.length)) {
+                            cb();
+                        }
+                        if (done) {
+                            cb();
+                            this.removeListener("chunk", pull);
+                            p.end();
+                            return;
+                        }
+                        // Continue draining regardless of downstream read timing.
+                        queueMicrotask(pull);
+                    });
+                    return;
+                }
+            }
+            if (!done) {
+                if (this.finished) {
+                    this.removeListener("chunk", pull);
+                    cb();
+                    p.destroy(new Error("FILE_ENDED"));
+                    return;
+                }
+            }
+            else {
+                this.removeListener("chunk", pull);
+                cb();
+                p.end();
+            }
+        };
+        this.on("chunk", pull);
+        pull();
+        return p;
+    }
+    pull(eof, includeEof) {
+        if (eof === 0) {
+            return Promise.resolve(new Uint8Array(0));
+        }
+        // If we already have the required data in buffer
+        // we can resolve the request immediately
+        if (typeof eof === "number" && this._queue.length >= eof) {
+            const data = this._queue.read(eof);
+            // If we drained the internal buffer, allow the upstream writer to continue.
+            if (this._queue.length === 0) {
+                this._maybeReleaseWriteCallback();
+            }
+            return Promise.resolve(data);
+        }
+        // Otherwise we stream until we have it
+        const chunks = [];
+        const concatStream = new Transform({
+            transform(d, _encoding, cb) {
+                chunks.push(d);
+                cb();
+            }
+        });
+        let pullStreamRejectHandler;
+        return new Promise((resolve, reject) => {
+            pullStreamRejectHandler = (e) => {
+                this.__emittedError = e;
+                reject(e);
+            };
+            if (this.finished) {
+                return reject(new Error("FILE_ENDED"));
+            }
+            this.once("error", pullStreamRejectHandler); // reject any errors from pullstream itself
+            this.stream(eof, includeEof)
+                .on("error", reject)
+                .pipe(concatStream)
+                .on("finish", () => {
+                resolve(chunks.length === 1 ? chunks[0] : concatUint8Arrays(chunks));
+            })
+                .on("error", reject);
+        }).finally(() => {
+            this.removeListener("error", pullStreamRejectHandler);
+        });
+    }
+    pullUntil(pattern, includeEof) {
+        return this.pull(pattern, includeEof);
+    }
+}
+export async function readCrxHeader(pull) {
+    const data = await pull(12);
+    const header = parseBuffer(data, CRX_HEADER_FORMAT);
+    const pubKeyLength = header.pubKeyLength || 0;
+    const signatureLength = header.signatureLength || 0;
+    const keyAndSig = await pull(pubKeyLength + signatureLength);
+    header.publicKey = keyAndSig.subarray(0, pubKeyLength);
+    header.signature = keyAndSig.subarray(pubKeyLength);
+    return header;
+}
+export async function readLocalFileHeader(pull) {
+    const data = await pull(26);
+    const vars = parseBuffer(data, LOCAL_FILE_HEADER_FORMAT);
+    const fileNameBuffer = await pull(vars.fileNameLength || 0);
+    const extraFieldData = await pull(vars.extraFieldLength || 0);
+    return { vars, fileNameBuffer, extraFieldData };
+}
+export async function readDataDescriptor(pull) {
+    const data = await pull(16);
+    return parseBuffer(data, DATA_DESCRIPTOR_FORMAT);
+}
+export async function consumeCentralDirectoryFileHeader(pull) {
+    const data = await pull(42);
+    const vars = parseBuffer(data, CENTRAL_DIRECTORY_FILE_HEADER_FORMAT);
+    await pull(vars.fileNameLength || 0);
+    await pull(vars.extraFieldLength || 0);
+    await pull(vars.fileCommentLength || 0);
+}
+export async function consumeEndOfCentralDirectoryRecord(pull) {
+    const data = await pull(18);
+    const vars = parseBuffer(data, END_OF_CENTRAL_DIRECTORY_FORMAT);
+    await pull(vars.commentLength || 0);
+}
+// =============================================================================
+// Validated Data Descriptor Scan (shared by Node + Browser)
+// =============================================================================
+function isValidZipRecordSignature(sig) {
+    switch (sig) {
+        case LOCAL_FILE_HEADER_SIG:
+        case CENTRAL_DIR_HEADER_SIG:
+        case END_OF_CENTRAL_DIR_SIG:
+        case ZIP64_END_OF_CENTRAL_DIR_SIG:
+        case ZIP64_END_OF_CENTRAL_DIR_LOCATOR_SIG:
+            return true;
+        default:
+            return false;
+    }
+}
+function readUint32LEFromBytes(view, offset) {
+    return ((view[offset] |
+        0 |
+        ((view[offset + 1] | 0) << 8) |
+        ((view[offset + 2] | 0) << 16) |
+        ((view[offset + 3] | 0) << 24)) >>>
+        0);
+}
+function indexOf4BytesPattern(buffer, pattern, startIndex) {
+    if (pattern.length !== 4) {
+        return indexOfUint8ArrayPattern(buffer, pattern, startIndex);
+    }
+    const b0 = pattern[0];
+    const b1 = pattern[1];
+    const b2 = pattern[2];
+    const b3 = pattern[3];
+    const bufLen = buffer.length;
+    let start = startIndex | 0;
+    if (start < 0) {
+        start = 0;
+    }
+    if (start > bufLen - 4) {
+        return -1;
+    }
+    for (let i = start; i <= bufLen - 4; i++) {
+        if (buffer[i] === b0 && buffer[i + 1] === b1 && buffer[i + 2] === b2 && buffer[i + 3] === b3) {
+            return i;
+        }
+    }
+    return -1;
+}
+function initScanResult(out) {
+    if (out) {
+        return out;
+    }
+    return { foundIndex: -1, nextSearchFrom: 0 };
+}
+/**
+ * Scan for a validated DATA_DESCRIPTOR record boundary.
+ *
+ * Scanning for the 4-byte signature alone is unsafe because it can appear inside
+ * compressed data. We validate a candidate by requiring:
+ * - the next 4 bytes after the 16-byte descriptor form a known ZIP record signature, and
+ * - the descriptor's compressedSize matches the number of compressed bytes emitted so far.
+ */
+export function scanValidatedDataDescriptor(view, dataDescriptorSignature, bytesEmitted, startIndex = 0, out) {
+    const result = initScanResult(out);
+    let searchFrom = startIndex | 0;
+    if (searchFrom < 0) {
+        searchFrom = 0;
+    }
+    if (searchFrom > view.length) {
+        searchFrom = view.length;
+    }
+    // To avoid missing a signature split across chunk boundaries, we may need
+    // to re-check the last (sigLen - 1) bytes on the next scan.
+    const sigLen = dataDescriptorSignature.length | 0;
+    const overlap = sigLen > 0 ? sigLen - 1 : 0;
+    while (searchFrom < view.length) {
+        const match = indexOf4BytesPattern(view, dataDescriptorSignature, searchFrom);
+        if (match === -1) {
+            result.foundIndex = -1;
+            result.nextSearchFrom = Math.max(searchFrom, Math.max(0, view.length - overlap));
+            return result;
+        }
+        const idx = match;
+        // Need 16 bytes for descriptor + 4 bytes for next record signature.
+        const nextSigOffset = idx + 16;
+        if (nextSigOffset + 4 <= view.length) {
+            const nextSig = readUint32LEFromBytes(view, nextSigOffset);
+            const descriptorCompressedSize = readUint32LEFromBytes(view, idx + 8);
+            const expectedCompressedSize = (bytesEmitted + idx) >>> 0;
+            if (isValidZipRecordSignature(nextSig) &&
+                descriptorCompressedSize === expectedCompressedSize) {
+                result.foundIndex = idx;
+                result.nextSearchFrom = idx;
+                return result;
+            }
+            searchFrom = idx + 1;
+            continue;
+        }
+        // Not enough bytes to validate yet. Re-check this candidate once more bytes arrive.
+        result.foundIndex = -1;
+        result.nextSearchFrom = idx;
+        return result;
+    }
+    result.foundIndex = -1;
+    result.nextSearchFrom = Math.max(searchFrom, Math.max(0, view.length - overlap));
+    return result;
+}
+/**
+ * Stream compressed file data until we reach a validated DATA_DESCRIPTOR boundary.
+ *
+ * This encapsulates the shared logic used by both Node and browser parsers.
+ */
+export function streamUntilValidatedDataDescriptor(options) {
+    const { source, dataDescriptorSignature } = options;
+    const keepTailBytes = options.keepTailBytes ?? 20;
+    const errorMessage = options.errorMessage ?? "FILE_ENDED: Data descriptor not found";
+    const output = new PassThrough();
+    let done = false;
+    // Total number of compressed bytes already emitted for this entry.
+    let bytesEmitted = 0;
+    let searchFrom = 0;
+    const scanResult = { foundIndex: -1, nextSearchFrom: 0 };
+    let unsubscribe;
+    const cleanup = () => {
+        if (unsubscribe) {
+            unsubscribe();
+            unsubscribe = undefined;
+        }
+    };
+    const pull = () => {
+        if (done) {
+            return;
+        }
+        while (source.getLength() > 0) {
+            const view = source.getView();
+            scanValidatedDataDescriptor(view, dataDescriptorSignature, bytesEmitted, searchFrom, scanResult);
+            const foundIndex = scanResult.foundIndex;
+            searchFrom = scanResult.nextSearchFrom;
+            if (foundIndex !== -1) {
+                if (foundIndex > 0) {
+                    output.write(source.read(foundIndex));
+                    bytesEmitted += foundIndex;
+                    searchFrom = Math.max(0, searchFrom - foundIndex);
+                }
+                done = true;
+                source.maybeReleaseWriteCallback?.();
+                cleanup();
+                output.end();
+                return;
+            }
+            // Flush most of the buffered data but keep a tail so a potential signature
+            // split across chunks can still be detected/validated.
+            const flushLen = Math.max(0, view.length - keepTailBytes);
+            if (flushLen > 0) {
+                output.write(source.read(flushLen));
+                bytesEmitted += flushLen;
+                searchFrom = Math.max(0, searchFrom - flushLen);
+                if (source.getLength() <= keepTailBytes) {
+                    source.maybeReleaseWriteCallback?.();
+                }
+                return;
+            }
+            // Need more data.
+            break;
+        }
+        if (!done && source.isFinished()) {
+            done = true;
+            cleanup();
+            output.destroy(new Error(errorMessage));
+        }
+    };
+    unsubscribe = source.onDataAvailable(pull);
+    queueMicrotask(pull);
+    return output;
+}
+const endDirectorySignature = writeUint32LE(END_OF_CENTRAL_DIR_SIG);
+export async function runParseLoop(opts, io, emitter, inflateFactory, state) {
+    while (true) {
+        const sigBytes = await io.pull(4);
+        if (sigBytes.length === 0) {
+            emitter.emitClose();
+            return;
+        }
+        const signature = readUint32LE(sigBytes, 0);
+        if (signature === 0x34327243) {
+            state.crxHeader = await readCrxHeader(async (length) => io.pull(length));
+            emitter.emitCrxHeader(state.crxHeader);
+            continue;
+        }
+        if (signature === LOCAL_FILE_HEADER_SIG) {
+            await readFileRecord(opts, io, emitter, inflateFactory, state);
+            continue;
+        }
+        if (signature === CENTRAL_DIR_HEADER_SIG) {
+            state.reachedCD = true;
+            await consumeCentralDirectoryFileHeader(async (length) => io.pull(length));
+            continue;
+        }
+        if (signature === END_OF_CENTRAL_DIR_SIG) {
+            await consumeEndOfCentralDirectoryRecord(async (length) => io.pull(length));
+            io.setDone();
+            emitter.emitClose();
+            return;
+        }
+        if (state.reachedCD) {
+            // We are in central directory trailing data; resync by scanning for EOCD signature.
+            // consumeEndOfCentralDirectoryRecord expects the EOCD signature to be consumed, so includeEof=true.
+            const includeEof = true;
+            await io.pullUntil(endDirectorySignature, includeEof);
+            await consumeEndOfCentralDirectoryRecord(async (length) => io.pull(length));
+            io.setDone();
+            emitter.emitClose();
+            return;
+        }
+        emitter.emitError(new Error("invalid signature: 0x" + signature.toString(16)));
+        emitter.emitClose();
+        return;
+    }
+}
+async function readFileRecord(opts, io, emitter, inflateFactory, state) {
+    const { vars: headerVars, fileNameBuffer, extraFieldData } = await readLocalFileHeader(async (l) => io.pull(l));
+    const vars = headerVars;
+    if (state.crxHeader) {
+        vars.crxHeader = state.crxHeader;
+    }
+    const fileName = decodeZipEntryPath(fileNameBuffer);
+    const entry = new PassThrough();
+    let autodraining = false;
+    entry.autodrain = function () {
+        autodraining = true;
+        entry.__autodraining = true;
+        return autodrain(entry);
+    };
+    entry.buffer = function () {
+        return bufferStream(entry);
+    };
+    entry.path = fileName;
+    entry.props = buildZipEntryProps(fileName, fileNameBuffer, vars.flags);
+    entry.type = getZipEntryType(fileName, vars.uncompressedSize || 0);
+    if (opts.verbose) {
+        if (entry.type === "Directory") {
+            console.log("   creating:", fileName);
+        }
+        else if (entry.type === "File") {
+            if (vars.compressionMethod === 0) {
+                console.log(" extracting:", fileName);
+            }
+            else {
+                console.log("  inflating:", fileName);
+            }
+        }
+    }
+    const extra = parseExtraField(extraFieldData, vars);
+    vars.lastModifiedDateTime = resolveZipEntryLastModifiedDateTime(vars, extra);
+    entry.vars = vars;
+    entry.extraFields = extra;
+    entry.__autodraining = autodraining;
+    const fileSizeKnown = isFileSizeKnown(vars.flags, vars.compressedSize);
+    if (fileSizeKnown) {
+        entry.size = vars.uncompressedSize || 0;
+    }
+    if (opts.forceStream) {
+        emitter.pushEntry(entry);
+    }
+    else {
+        emitter.emitEntry(entry);
+        emitter.pushEntryIfPiped(entry);
+    }
+    if (opts.verbose) {
+        console.log({
+            filename: fileName,
+            vars: vars,
+            extraFields: entry.extraFields
+        });
+    }
+    const inflater = vars.compressionMethod && !autodraining ? inflateFactory() : new PassThrough();
+    if (fileSizeKnown) {
+        await pipeline(io.stream(vars.compressedSize || 0), inflater, entry);
+        return;
+    }
+    await pipeline(io.streamUntilDataDescriptor(), inflater, entry);
+    const dd = await readDataDescriptor(async (l) => io.pull(l));
+    entry.size = dd.uncompressedSize || 0;
+}

package/dist/browser/modules/archive/parse.browser.d.ts

@@ -0,0 +1,22 @@
+/**
+ * ZIP Stream Parser - Browser Version
+ *
+ * A streaming ZIP parser for browsers using native DecompressionStream.
+ * Falls back to pure JavaScript implementation for older browsers.
+ * Uses the browser Duplex stream implementation for compatibility.
+ */
+import { Duplex } from "@stream";
+import { type CrxHeader, type PullStreamPublicApi, type EntryProps, type EntryVars, type InflateFactory, type ParseOptions, type ZipEntry } from "@archive/parse.base";
+export type { CrxHeader, EntryProps, EntryVars, ParseOptions, ZipEntry };
+export type ParseStream = Duplex & {
+    promise(): Promise<void>;
+} & PullStreamPublicApi & {
+    crxHeader?: CrxHeader;
+};
+export declare function createParseClass(createInflateRawFn: InflateFactory): {
+    new (opts?: ParseOptions): ParseStream;
+};
+declare const BaseParse: new (opts?: ParseOptions) => ParseStream;
+export declare class Parse extends BaseParse {
+}
+export declare function createParse(opts?: ParseOptions): ParseStream;