berstend / unzip-bug

Meta

crx is the file format of chrome extensions. It's basically zip with some extra stuff in the beginning.

Goal: Fetch crx file, strip crx headers on the fly, pipe it to node-unzip, extract only contents of interest.

---

Piping a crx directly to node-unzip:

~/d/t/unzip-bug ❯❯❯ coffee bug.coffee
stream.js:94
      throw er; // Unhandled stream error in pipe.
            ^
Error: invalid signature: 0x34327243
  at /Users/user/dev/testing/unzip-bug/node_modules/unzip/lib/parse.js:63:13
  at process._tickCallback (node.js:343:11)

node-unzip @parse.js

var signature = data.readUInt32LE(0);
if (signature === 0x04034b50) {
  self._readFile();
} else if (signature === 0x02014b50) {
  self._readCentralDirectoryFileHeader();
} else if (signature === 0x06054b50) {
  self._readEndOfCentralDirectoryRecord();
} else {
  err = new Error('invalid signature: 0x' + signature.toString(16));
  self.emit('error', err);
}

Slicing away the first 306 bytes off the first chunk results in true .zip

~/d/t/unzip-bug ❯❯❯ coffee bug.coffee
Signature pre transform 34327243
Signature post transform 4034b50
128.png File undefined
events.js:85
      throw er; // Unhandled 'error' event
            ^
Error: invalid signature: 0x8080014
  at /Users/user/dev/testing/unzip-bug/node_modules/unzip/lib/parse.js:63:13
  at process._tickCallback (node.js:343:11)

4034b50 is a correct .zip

For some reason node-unzip still receives the wrong or corrupt chunk.

Testing

npm install
coffee bug.coffee