You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
323 lines
9.8 KiB
323 lines
9.8 KiB
'use strict';
|
|
|
|
// The zip file spec is at http://www.pkware.com/documents/casestudies/APPNOTE.TXT
|
|
// TODO: There is fair chunk of the spec that I have ignored. Need to add
|
|
// assertions everywhere to make sure that we are not dealing with a ZIP type
|
|
// that I haven't designed for. Things like spanning archives, non-DEFLATE
|
|
// compression, encryption, etc.
|
|
var fs = require('graceful-fs');
|
|
var Q = require('q');
|
|
var path = require('path');
|
|
var util = require('util');
|
|
var events = require('events');
|
|
var structures = require('./structures');
|
|
var signatures = require('./signatures');
|
|
var extractors = require('./extractors');
|
|
var FileDetails = require('./file-details');
|
|
|
|
var fstat = Q.denodeify(fs.fstat);
|
|
var read = Q.denodeify(fs.read);
|
|
var fopen = Q.denodeify(fs.open);
|
|
|
|
function DecompressZip(filename) {
|
|
events.EventEmitter.call(this);
|
|
|
|
this.filename = filename;
|
|
this.stats = null;
|
|
this.fd = null;
|
|
this.chunkSize = 1024 * 1024; // Buffer up to 1Mb at a time
|
|
this.dirCache = {};
|
|
|
|
// When we need a resource, we should check if there is a promise for it
|
|
// already and use that. If the promise is already fulfilled we don't do the
|
|
// async work again and we get to queue up dependant tasks.
|
|
this._p = {}; // _p instead of _promises because it is a lot easier to read
|
|
}
|
|
|
|
util.inherits(DecompressZip, events.EventEmitter);
|
|
|
|
DecompressZip.prototype.openFile = function () {
|
|
return fopen(this.filename, 'r');
|
|
};
|
|
|
|
DecompressZip.prototype.closeFile = function () {
|
|
if (this.fd) {
|
|
fs.closeSync(this.fd);
|
|
this.fd = null;
|
|
}
|
|
};
|
|
|
|
DecompressZip.prototype.statFile = function (fd) {
|
|
this.fd = fd;
|
|
return fstat(fd);
|
|
};
|
|
|
|
DecompressZip.prototype.list = function () {
|
|
var self = this;
|
|
|
|
this.getFiles()
|
|
.then(function (files) {
|
|
var result = [];
|
|
|
|
files.forEach(function (file) {
|
|
result.push(file.path);
|
|
});
|
|
|
|
self.emit('list', result);
|
|
})
|
|
.fail(function (error) {
|
|
self.emit('error', error);
|
|
})
|
|
.fin(self.closeFile.bind(self));
|
|
|
|
return this;
|
|
};
|
|
|
|
DecompressZip.prototype.extract = function (options) {
|
|
var self = this;
|
|
|
|
options = options || {};
|
|
options.path = options.path || process.cwd();
|
|
options.filter = options.filter || null;
|
|
options.follow = !!options.follow;
|
|
options.strip = +options.strip || 0;
|
|
options.restrict = options.restrict !== false;
|
|
|
|
|
|
this.getFiles()
|
|
.then(function (files) {
|
|
var copies = [];
|
|
if (options.restrict) {
|
|
files = files.map(function (file) {
|
|
var destination = path.join(options.path, file.path);
|
|
// The destination path must not be outside options.path
|
|
if (destination.indexOf(options.path) !== 0) {
|
|
throw new Error('You cannot extract a file outside of the target path');
|
|
}
|
|
return file;
|
|
});
|
|
}
|
|
if (options.filter) {
|
|
files = files.filter(options.filter);
|
|
}
|
|
|
|
if (options.follow) {
|
|
copies = files.filter(function (file) {
|
|
return file.type === 'SymbolicLink';
|
|
});
|
|
files = files.filter(function (file) {
|
|
return file.type !== 'SymbolicLink';
|
|
});
|
|
}
|
|
|
|
if (options.strip) {
|
|
files = files.map(function (file) {
|
|
if (file.type !== 'Directory') {
|
|
// we don't use `path.sep` as we're using `/` in Windows too
|
|
var dir = file.parent.split('/');
|
|
var filename = file.filename;
|
|
|
|
if (options.strip > dir.length) {
|
|
throw new Error('You cannot strip more levels than there are directories');
|
|
} else {
|
|
dir = dir.slice(options.strip);
|
|
}
|
|
|
|
file.path = path.join(dir.join(path.sep), filename);
|
|
return file;
|
|
}
|
|
});
|
|
}
|
|
|
|
return self.extractFiles(files, options)
|
|
.then(self.extractFiles.bind(self, copies, options));
|
|
})
|
|
.then(function (results) {
|
|
self.emit('extract', results);
|
|
})
|
|
.fail(function (error) {
|
|
self.emit('error', error);
|
|
})
|
|
.fin(self.closeFile.bind(self));
|
|
|
|
return this;
|
|
};
|
|
|
|
// Utility methods
|
|
DecompressZip.prototype.getSearchBuffer = function (stats) {
|
|
var size = Math.min(stats.size, this.chunkSize);
|
|
this.stats = stats;
|
|
return this.getBuffer(stats.size - size, stats.size);
|
|
};
|
|
|
|
DecompressZip.prototype.getBuffer = function (start, end) {
|
|
var size = end - start;
|
|
return read(this.fd, new Buffer(size), 0, size, start)
|
|
.then(function (result) {
|
|
return result[1];
|
|
});
|
|
};
|
|
|
|
DecompressZip.prototype.findEndOfDirectory = function (buffer) {
|
|
var index = buffer.length - 3;
|
|
var chunk = '';
|
|
|
|
// Apparently the ZIP spec is not very good and it is impossible to
|
|
// guarantee that you have read a zip file correctly, or to determine
|
|
// the location of the CD without hunting.
|
|
// Search backwards through the buffer, as it is very likely to be near the
|
|
// end of the file.
|
|
while (index > Math.max(buffer.length - this.chunkSize, 0) && chunk !== signatures.END_OF_CENTRAL_DIRECTORY) {
|
|
index--;
|
|
chunk = buffer.readUInt32LE(index);
|
|
}
|
|
|
|
if (chunk !== signatures.END_OF_CENTRAL_DIRECTORY) {
|
|
throw new Error('Could not find the End of Central Directory Record');
|
|
}
|
|
|
|
return buffer.slice(index);
|
|
};
|
|
|
|
// Directory here means the ZIP Central Directory, not a folder
|
|
DecompressZip.prototype.readDirectory = function (recordBuffer) {
|
|
var record = structures.readEndRecord(recordBuffer);
|
|
|
|
return this.getBuffer(record.directoryOffset, record.directoryOffset + record.directorySize)
|
|
.then(structures.readDirectory.bind(null));
|
|
};
|
|
|
|
DecompressZip.prototype.getFiles = function () {
|
|
if (!this._p.getFiles) {
|
|
this._p.getFiles = this.openFile()
|
|
.then(this.statFile.bind(this))
|
|
.then(this.getSearchBuffer.bind(this))
|
|
.then(this.findEndOfDirectory.bind(this))
|
|
.then(this.readDirectory.bind(this))
|
|
.then(this.readFileEntries.bind(this));
|
|
}
|
|
|
|
return this._p.getFiles;
|
|
};
|
|
|
|
DecompressZip.prototype.readFileEntries = function (directory) {
|
|
var promises = [];
|
|
var files = [];
|
|
var self = this;
|
|
|
|
directory.forEach(function (directoryEntry, index) {
|
|
var start = directoryEntry.relativeOffsetOfLocalHeader;
|
|
var end = Math.min(self.stats.size, start + structures.maxFileEntrySize);
|
|
var fileDetails = new FileDetails(directoryEntry);
|
|
|
|
var promise = self.getBuffer(start, end)
|
|
.then(structures.readFileEntry.bind(null))
|
|
.then(function (fileEntry) {
|
|
var maxSize;
|
|
|
|
if (fileDetails.compressedSize > 0) {
|
|
maxSize = fileDetails.compressedSize;
|
|
} else {
|
|
maxSize = self.stats.size;
|
|
|
|
if (index < directory.length - 1) {
|
|
maxSize = directory[index + 1].relativeOffsetOfLocalHeader;
|
|
}
|
|
|
|
maxSize -= start + fileEntry.entryLength;
|
|
}
|
|
|
|
fileDetails._offset = start + fileEntry.entryLength;
|
|
fileDetails._maxSize = maxSize;
|
|
|
|
self.emit('file', fileDetails);
|
|
files[index] = fileDetails;
|
|
});
|
|
|
|
promises.push(promise);
|
|
});
|
|
|
|
return Q.all(promises)
|
|
.then(function () {
|
|
return files;
|
|
});
|
|
};
|
|
|
|
DecompressZip.prototype.extractFiles = function (files, options, results) {
|
|
var promises = [];
|
|
var self = this;
|
|
|
|
results = results || [];
|
|
var fileIndex = 0;
|
|
files.forEach(function (file) {
|
|
var promise = self.extractFile(file, options)
|
|
.then(function (result) {
|
|
self.emit('progress', fileIndex++, files.length);
|
|
results.push(result);
|
|
});
|
|
|
|
promises.push(promise);
|
|
});
|
|
|
|
return Q.all(promises)
|
|
.then(function () {
|
|
return results;
|
|
});
|
|
};
|
|
|
|
DecompressZip.prototype.extractFile = function (file, options) {
|
|
var destination = path.join(options.path, file.path);
|
|
|
|
// Possible compression methods:
|
|
// 0 - The file is stored (no compression)
|
|
// 1 - The file is Shrunk
|
|
// 2 - The file is Reduced with compression factor 1
|
|
// 3 - The file is Reduced with compression factor 2
|
|
// 4 - The file is Reduced with compression factor 3
|
|
// 5 - The file is Reduced with compression factor 4
|
|
// 6 - The file is Imploded
|
|
// 7 - Reserved for Tokenizing compression algorithm
|
|
// 8 - The file is Deflated
|
|
// 9 - Enhanced Deflating using Deflate64(tm)
|
|
// 10 - PKWARE Data Compression Library Imploding (old IBM TERSE)
|
|
// 11 - Reserved by PKWARE
|
|
// 12 - File is compressed using BZIP2 algorithm
|
|
// 13 - Reserved by PKWARE
|
|
// 14 - LZMA (EFS)
|
|
// 15 - Reserved by PKWARE
|
|
// 16 - Reserved by PKWARE
|
|
// 17 - Reserved by PKWARE
|
|
// 18 - File is compressed using IBM TERSE (new)
|
|
// 19 - IBM LZ77 z Architecture (PFS)
|
|
// 97 - WavPack compressed data
|
|
// 98 - PPMd version I, Rev 1
|
|
|
|
if (file.type === 'Directory') {
|
|
return extractors.folder(file, destination, this);
|
|
}
|
|
|
|
if (file.type === 'File') {
|
|
switch (file.compressionMethod) {
|
|
case 0:
|
|
return extractors.store(file, destination, this);
|
|
|
|
case 8:
|
|
return extractors.deflate(file, destination, this);
|
|
|
|
default:
|
|
throw new Error('Unsupported compression type');
|
|
}
|
|
}
|
|
|
|
if (file.type === 'SymbolicLink') {
|
|
if (options.follow) {
|
|
return extractors.copy(file, destination, this, options.path);
|
|
} else {
|
|
return extractors.symlink(file, destination, this, options.path);
|
|
}
|
|
}
|
|
|
|
throw new Error('Unsupported file type "' + file.type + '"');
|
|
};
|
|
|
|
module.exports = DecompressZip;
|
|
|