gzip-js/lib/gzip.js

GZIP in pure JavaScript (works in the browser)

(function () {
	'use strict';

	var crc32 = require('crc32'),
		deflate = require('deflate-js'),
		// magic numbers marking this file as GZIP
		ID1 = 0x1F,
		ID2 = 0x8B,
		compressionMethods = {
			'deflate': 8
		},
		possibleFlags = {
			'FTEXT': 0x01,
			'FHCRC': 0x02,
			'FEXTRA': 0x04,
			'FNAME': 0x08,
			'FCOMMENT': 0x10
		},
		osMap = {
			'fat': 0, // FAT file system (DOS, OS/2, NT) + PKZIPW 2.50 VFAT, NTFS
			'amiga': 1, // Amiga
			'vmz': 2, // VMS (VAX or Alpha AXP)
			'unix': 3, // Unix
			'vm/cms': 4, // VM/CMS
			'atari': 5, // Atari
			'hpfs': 6, // HPFS file system (OS/2, NT 3.x)
			'macintosh': 7, // Macintosh
			'z-system': 8, // Z-System
			'cplm': 9, // CP/M
			'tops-20': 10, // TOPS-20
			'ntfs': 11, // NTFS file system (NT)
			'qdos': 12, // SMS/QDOS
			'acorn': 13, // Acorn RISC OS
			'vfat': 14, // VFAT file system (Win95, NT)
			'vms': 15, // MVS (code also taken for PRIMOS)
			'beos': 16, // BeOS (BeBox or PowerMac)
			'tandem': 17, // Tandem/NSK
			'theos': 18 // THEOS
		},
		os = 'unix',
		DEFAULT_LEVEL = 6;

	function putByte(n, arr) {
		arr.push(n & 0xFF);
	}

	// LSB first
	function putShort(n, arr) {
		arr.push(n & 0xFF);
		arr.push(n >>> 8);
	}

	// LSB first
	function putLong(n, arr) {
		putShort(n & 0xffff, arr);
		putShort(n >>> 16, arr);
	}

	function putString(s, arr) {
		var i, len = s.length;
		for (i = 0; i < len; i += 1) {
			putByte(s.charCodeAt(i), arr);
		}
	}

	function readByte(arr) {
		return arr.shift();
	}

	function readShort(arr) {
		return arr.shift() | (arr.shift() << 8);
	}

	function readLong(arr) {
		var n1 = readShort(arr),
			n2 = readShort(arr);

		// JavaScript can't handle bits in the position 32
		// we'll emulate this by removing the left-most bit (if it exists)
		// and add it back in via multiplication, which does work
		if (n2 > 32768) {
			n2 -= 32768;

			return ((n2 << 16) | n1) + 32768 * Math.pow(2, 16);
		}

		return (n2 << 16) | n1;
	}

	function readString(arr) {
		var charArr = [];

		// turn all bytes into chars until the terminating null
		while (arr[0] !== 0) {
			charArr.push(String.fromCharCode(arr.shift()));
		}

		// throw away terminating null
		arr.shift();

		// join all characters into a cohesive string
		return charArr.join('');
	}

	/*
	 * Reads n number of bytes and return as an array.
	 *
	 * @param arr- Array of bytes to read from
	 * @param n- Number of bytes to read
	 */
	function readBytes(arr, n) {
		var i, ret = [];
		for (i = 0; i < n; i += 1) {
			ret.push(arr.shift());
		}

		return ret;
	}

	/*
	 * ZIPs a file in GZIP format. The format is as given by the spec, found at:
	 * http://www.gzip.org/zlib/rfc-gzip.html
	 *
	 * Omitted parts in this implementation:
	 */
	function zip(data, options) {
		var flags = 0,
			level,
			crc, out = [];

		if (!options) {
			options = {};
		}
		level = options.level || DEFAULT_LEVEL;

		if (typeof data === 'string') {
			data = Array.prototype.map.call(data, function (char) {
				return char.charCodeAt(0);
			});
		}

		// magic number marking this file as GZIP
		putByte(ID1, out);
		putByte(ID2, out);

		putByte(compressionMethods['deflate'], out);

		if (options.name) {
			flags |= possibleFlags['FNAME'];
		}

		putByte(flags, out);
		putLong(options.timestamp || parseInt(Date.now() / 1000, 10), out);

		// put deflate args (extra flags)
		if (level === 1) {
			// fastest algorithm
			putByte(4, out);
		} else if (level === 9) {
			// maximum compression (fastest algorithm)
			putByte(2, out);
		} else {
			putByte(0, out);
		}

		// OS identifier
		putByte(osMap[os], out);

		if (options.name) {
			// ignore the directory part
			putString(options.name.substring(options.name.lastIndexOf('/') + 1), out);

			// terminating null
			putByte(0, out);
		}

		deflate.deflate(data, level).forEach(function (byte) {
			putByte(byte, out);
		});

		putLong(parseInt(crc32(data), 16), out);
		putLong(data.length, out);

		return out;
	}

	function unzip(data, options) {
		// start with a copy of the array
		var arr = Array.prototype.slice.call(data, 0),
			t,
			compressionMethod,
			flags,
			mtime,
			xFlags,
			key,
			os,
			crc,
			size,
			res;

		// check the first two bytes for the magic numbers
		if (readByte(arr) !== ID1 || readByte(arr) !== ID2) {
			throw 'Not a GZIP file';
		}

		t = readByte(arr);
		t = Object.keys(compressionMethods).some(function (key) {
			compressionMethod = key;
			return compressionMethods[key] === t;
		});

		if (!t) {
			throw 'Unsupported compression method';
		}

		flags = readByte(arr);
		mtime = readLong(arr);
		xFlags = readByte(arr);
		t = readByte(arr);
		Object.keys(osMap).some(function (key) {
			if (osMap[key] === t) {
				os = key;
				return true;
			}
		});

		// just throw away the bytes for now
		if (flags & possibleFlags['FEXTRA']) {
			t = readShort(arr);
			readBytes(arr, t);
		}

		// just throw away for now
		if (flags & possibleFlags['FNAME']) {
			readString(arr);
		}

		// just throw away for now
		if (flags & possibleFlags['FCOMMENT']) {
			readString(arr);
		}

		// just throw away for now
		if (flags & possibleFlags['FHCRC']) {
			readShort(arr);
		}

		if (compressionMethod === 'deflate') {
			// give deflate everything but the last 8 bytes
			// the last 8 bytes are for the CRC32 checksum and filesize
			res = deflate.inflate(arr.splice(0, arr.length - 8));
		}

		if (flags & possibleFlags['FTEXT']) {
			res = Array.prototype.map.call(res, function (byte) {
				return String.fromCharCode(byte);
			}).join('');
		}

		crc = readLong(arr);
		if (crc !== parseInt(crc32(res), 16)) {
			throw 'Checksum does not match';
		}

		size = readLong(arr);
		if (size !== res.length) {
			throw 'Size of decompressed file not correct';
		}

		return res;
	}

	module.exports = {
		zip: zip,
		unzip: unzip,
		get DEFAULT_LEVEL() {
			return DEFAULT_LEVEL;
		}
	};
}());