Fix issue #38: Use runtime's native DecompressionStream to inflate zip streams.

CHANGELOG.md

@@ -7,9 +7,16 @@ All notable changes to this project will be documented in this file.
 ### Added
 
 - archive: Support DEFLATE in Zipper where JS implementations support it in CompressionStream.
+  [Issue #40](https://github.com/codedread/bitjs/issues/40)
+- archive: Support DEFLATE in Unzipper where JS implementations support it in DecompressionStream.
+  [Issue #38](https://github.com/codedread/bitjs/issues/38)
 - file: Added detection of GZIP files.
 - io: Added a skip() method to BitStream to match ByteStream.
 
+### Fixed
+
+- Fixed a benign JS error in the Web Worker wrapper
+
 ## [1.2.1] - 2024-01-19
 
 ### Added

README.md

@@ -8,15 +8,13 @@ A set of dependency-free JavaScript modules to handle binary data in JS (using
 [Typed Arrays](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/TypedArray)).
 Includes:
 
-  * bitjs/archive: Unarchiving files (unzip, unrar, untar) in JavaScript,
-    implemented as Web Workers where supported, and allowing progressive
-    unarchiving while streaming.
-  * bitjs/codecs: Get the codec info of media containers in a ISO RFC6381
-    MIME type string
+  * bitjs/archive: Decompressing files (unzip, unrar, untar) in JavaScript, implemented as Web
+    Workers where supported, and allowing progressive unarchiving while streaming.
+  * bitjs/codecs: Get the codec info of media containers in a ISO RFC6381 MIME type string.
   * bitjs/file: Detect the type of file from its binary signature.
   * bitjs/image: Parsing GIF, JPEG, PNG. Conversion of WebP to PNG or JPEG.
-  * bitjs/io: Low-level classes for interpreting binary data (BitStream
-    ByteStream).  For example, reading or peeking at N bits at a time.
+  * bitjs/io: Low-level classes for interpreting binary data (BitStream, ByteStream). For example,
+    reading or peeking at N bits at a time.
 
 ## Installation
 

archive/decompress.js

@@ -316,3 +316,22 @@ export function getUnarchiver(ab, options = {}) {
   }
   return unarchiver;
 }
+
+// import * as fs from 'node:fs';
+// async function main() {
+//   const nodeBuf = fs.readFileSync(`./action-1.cbz`);
+//   const ab = nodeBuf.buffer.slice(nodeBuf.byteOffset, nodeBuf.byteOffset + nodeBuf.length);
+//   const then = Date.now();
+//   const zipper = new Unzipper(ab, {debug: true});
+//   zipper.addEventListener('extract', evt => {
+//     const f = evt.unarchivedFile;
+//     fs.writeFileSync(f.filename, Buffer.from(f.fileData));
+//   });
+//   zipper.addEventListener('finish', evt => {
+//     console.dir(evt);
+//     console.log(`Took ${(Date.now() - then)}ms`);
+//   });
+//   await zipper.start();
+// }
+
+// main();

archive/unzip.js

@@ -178,7 +178,7 @@ class ZipLocalFile {
   }
 
   // determine what kind of compressed data we have and decompress
-  unzip() {
+  async unzip() {
     if (!this.fileData) {
       err('unzip() called on a file with out compressed file data');
     }
@@ -196,7 +196,7 @@ class ZipLocalFile {
       if (logToConsole) {
         info(`ZIP v2.0, DEFLATE: ${this.filename} (${this.compressedSize} bytes)`);
       }
-      this.fileData = inflate(this.fileData, this.uncompressedSize);
+      this.fileData = await inflate(this.fileData, this.uncompressedSize);
     }
     else {
       err(`UNSUPPORTED VERSION/FORMAT: ZIP v${this.version}, ` +
@@ -483,9 +483,18 @@ function inflateBlockData(bstream, hcLiteralTable, hcDistanceTable, buffer) {
  * Compression method 8. Deflate: http://tools.ietf.org/html/rfc1951
  * @param {Uint8Array} compressedData A Uint8Array of the compressed file data.
  * @param {number} numDecompressedBytes
- * @returns {Uint8Array} The decompressed array.
+ * @returns {Promise<Uint8Array>} The decompressed array.
  */
-function inflate(compressedData, numDecompressedBytes) {
+async function inflate(compressedData, numDecompressedBytes) {
+  // Try to use native implementation of DEFLATE if it exists.
+  try {
+    const blob = new Blob([compressedData.buffer]);
+    const decompressedStream = blob.stream().pipeThrough(new DecompressionStream('deflate-raw'));
+    return new Uint8Array(await new Response(decompressedStream).arrayBuffer());
+  } catch (err) {
+    // Fall through to non-native implementation of DEFLATE.
+  }
+  
   // Bit stream representing the compressed data.
   /** @type {BitStream} */
   const bstream = new BitStream(compressedData.buffer,
@@ -596,7 +605,7 @@ function inflate(compressedData, numDecompressedBytes) {
   return buffer.data;
 }
 
-function archiveUnzip() {
+async function archiveUnzip() {
   let bstream = bytestream.tee();
 
   // loop until we don't see any more local files or we find a data descriptor.
@@ -619,7 +628,7 @@ function archiveUnzip() {
       currentBytesUnarchivedInFile = 0;
 
       // Actually do the unzipping.
-      oneLocalFile.unzip();
+      await oneLocalFile.unzip();
 
       if (oneLocalFile.fileData != null) {
         hostPort.postMessage({ type: 'extract', unarchivedFile: oneLocalFile }, [oneLocalFile.fileData.buffer]);
@@ -724,7 +733,7 @@ function archiveUnzip() {
 
 // event.data.file has the first ArrayBuffer.
 // event.data.bytes has all subsequent ArrayBuffers.
-const onmessage = function (event) {
+const onmessage = async function (event) {
   const bytes = event.data.file || event.data.bytes;
   logToConsole = !!event.data.logToConsole;
 
@@ -755,7 +764,7 @@ const onmessage = function (event) {
   if (unarchiveState === UnarchiveState.UNARCHIVING ||
     unarchiveState === UnarchiveState.WAITING) {
     try {
-      archiveUnzip();
+      await archiveUnzip();
     } catch (e) {
       if (typeof e === 'string' && e.startsWith('Error!  Overflowed')) {
         // Overrun the buffer.