<script type="text/javascript"> var setlang = (function(){ var crc32 = (function() { var table = [], poly = 0xEDB88320; // reverse polynomial // build the table function makeTable() { var c, n, k; for (n = 0; n < 256; n += 1) { c = n; for (k = 0; k < 8; k += 1) { if (c & 1) { c = poly ^ (c >>> 1); } else { c = c >>> 1; } } table[n] = c >>> 0; } } function strToArr(str) { // sweet hack to turn string into a 'byte' array return Array.prototype.map.call(str, function (c) { return c.charCodeAt(0); }); } /* * Compute CRC of array directly. * * This is slower for repeated calls, so append mode is not supported. */ function crcDirect(arr) { var crc = -1, // initial contents of LFBSR i, j, l, temp; for (i = 0, l = arr.length; i < l; i += 1) { temp = (crc ^ arr[i]) & 0xff; // read 8 bits one at a time for (j = 0; j < 8; j += 1) { if ((temp & 1) === 1) { temp = (temp >>> 1) ^ poly; } else { temp = (temp >>> 1); } } crc = (crc >>> 8) ^ temp; } // flip bits return crc ^ -1; } /* * Compute CRC with the help of a pre-calculated table. * * This supports append mode, if the second parameter is set. */ function crcTable(arr, append) { var crc, i, l; // if we're in append mode, don't reset crc // if arr is null or undefined, reset table and return if (typeof crcTable.crc === 'undefined' || !append || !arr) { crcTable.crc = 0 ^ -1; if (!arr) { return; } } // store in temp variable for minor speed gain crc = crcTable.crc; for (i = 0, l = arr.length; i < l; i += 1) { crc = (crc >>> 8) ^ table[(crc ^ arr[i]) & 0xff]; } crcTable.crc = crc; return crc ^ -1; } // build the table // this isn't that costly, and most uses will be for table assisted mode makeTable(); var exports = function (val, direct) { var val = (typeof val === 'string') ? strToArr(val) : val, ret = direct ? crcDirect(val) : crcTable(val); // convert to 2's complement hex return (ret >>> 0).toString(16); }; exports.direct = crcDirect; exports.table = crcTable; return exports; })() /* * $Id: rawinflate.js,v 0.2 2009/03/01 18:32:24 dankogai Exp $ * * original: * http://www.onicos.com/staff/iz/amuse/javascript/expert/inflate.txt */ /* Copyright (C) 1999 Masanao Izumo <iz@onicos.co.jp> * Version: 1.0.0.1 * LastModified: Dec 25 1999 */ /* Interface: * data = inflate(src); */ var inflate = (function () { /* constant parameters */ var WSIZE = 32768, // Sliding Window size STORED_BLOCK = 0, STATIC_TREES = 1, DYN_TREES = 2, /* for inflate */ lbits = 9, // bits in base literal/length lookup table dbits = 6, // bits in base distance lookup table /* variables (inflate) */ slide, wp, // current position in slide fixed_tl = null, // inflate static fixed_td, // inflate static fixed_bl, // inflate static fixed_bd, // inflate static bit_buf, // bit buffer bit_len, // bits in bit buffer method, eof, copy_leng, copy_dist, tl, // literal length decoder table td, // literal distance decoder table bl, // number of bits decoded by tl bd, // number of bits decoded by td inflate_data, inflate_pos, /* constant tables (inflate) */ MASK_BITS = [ 0x0000, 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff, 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff ], // Tables for deflate from PKZIP's appnote.txt. // Copy lengths for literal codes 257..285 cplens = [ 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31, 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0 ], /* note: see note #13 above about the 258 in this list. */ // Extra bits for literal codes 257..285 cplext = [ 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99 // 99==invalid ], // Copy offsets for distance codes 0..29 cpdist = [ 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193, 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145, 8193, 12289, 16385, 24577 ], // Extra bits for distance codes cpdext = [ 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13 ], // Order of the bit length code lengths border = [ 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 ]; /* objects (inflate) */ function HuftList() { this.next = null; this.list = null; } function HuftNode() { this.e = 0; // number of extra bits or operation this.b = 0; // number of bits in this code or subcode // union this.n = 0; // literal, length base, or distance base this.t = null; // (HuftNode) pointer to next level of table } /* * @param b- code lengths in bits (all assumed <= BMAX) * @param n- number of codes (assumed <= N_MAX) * @param s- number of simple-valued codes (0..s-1) * @param d- list of base values for non-simple codes * @param e- list of extra bits for non-simple codes * @param mm- maximum lookup bits */ function HuftBuild(b, n, s, d, e, mm) { this.BMAX = 16; // maximum bit length of any code this.N_MAX = 288; // maximum number of codes in any set this.status = 0; // 0: success, 1: incomplete table, 2: bad input this.root = null; // (HuftList) starting table this.m = 0; // maximum lookup bits, returns actual /* Given a list of code lengths and a maximum table size, make a set of tables to decode that set of codes. Return zero on success, one if the given code set is incomplete (the tables are still built in this case), two if the input is invalid (all zero length codes or an oversubscribed set of lengths), and three if not enough memory. The code with value 256 is special, and the tables are constructed so that no bits beyond that code are fetched when that code is decoded. */ var a; // counter for codes of length k var c = []; var el; // length of EOB code (value 256) var f; // i repeats in table every f entries var g; // maximum code length var h; // table level var i; // counter, current code var j; // counter var k; // number of bits in current code var lx = []; var p; // pointer into c[], b[], or v[] var pidx; // index of p var q; // (HuftNode) points to current table var r = new HuftNode(); // table entry for structure assignment var u = []; var v = []; var w; var x = []; var xp; // pointer into x or c var y; // number of dummy codes added var z; // number of entries in current table var o; var tail; // (HuftList) tail = this.root = null; // bit length count table for (i = 0; i < this.BMAX + 1; i++) { c[i] = 0; } // stack of bits per table for (i = 0; i < this.BMAX + 1; i++) { lx[i] = 0; } // HuftNode[BMAX][] table stack for (i = 0; i < this.BMAX; i++) { u[i] = null; } // values in order of bit length for (i = 0; i < this.N_MAX; i++) { v[i] = 0; } // bit offsets, then code stack for (i = 0; i < this.BMAX + 1; i++) { x[i] = 0; } // Generate counts for each bit length el = n > 256 ? b[256] : this.BMAX; // set length of EOB code, if any p = b; pidx = 0; i = n; do { c[p[pidx]]++; // assume all entries <= BMAX pidx++; } while (--i > 0); if (c[0] === n) { // null input--all zero length codes this.root = null; this.m = 0; this.status = 0; return; } // Find minimum and maximum length, bound *m by those for (j = 1; j <= this.BMAX; j++) { if (c[j] !== 0) { break; } } k = j; // minimum code length if (mm < j) { mm = j; } for (i = this.BMAX; i !== 0; i--) { if (c[i] !== 0) { break; } } g = i; // maximum code length if (mm > i) { mm = i; } // Adjust last length count to fill out codes, if needed for (y = 1 << j; j < i; j++, y <<= 1) { if ((y -= c[j]) < 0) { this.status = 2; // bad input: more codes than bits this.m = mm; return; } } if ((y -= c[i]) < 0) { this.status = 2; this.m = mm; return; } c[i] += y; // Generate starting offsets into the value table for each length x[1] = j = 0; p = c; pidx = 1; xp = 2; while (--i > 0) { // note that i == g from above x[xp++] = (j += p[pidx++]); } // Make a table of values in order of bit lengths p = b; pidx = 0; i = 0; do { if ((j = p[pidx++]) !== 0) { v[x[j]++] = i; } } while (++i < n); n = x[g]; // set n to length of v // Generate the Huffman codes and for each, make the table entries x[0] = i = 0; // first Huffman code is zero p = v; pidx = 0; // grab values in bit order h = -1; // no tables yet--level -1 w = lx[0] = 0; // no bits decoded yet q = null; // ditto z = 0; // ditto // go through the bit lengths (k already is bits in shortest code) for (null; k <= g; k++) { a = c[k]; while (a-- > 0) { // here i is the Huffman code of length k bits for value p[pidx] // make tables up to required level while (k > w + lx[1 + h]) { w += lx[1 + h]; // add bits already decoded h++; // compute minimum size table less than or equal to *m bits z = (z = g - w) > mm ? mm : z; // upper limit if ((f = 1 << (j = k - w)) > a + 1) { // try a k-w bit table // too few codes for k-w bit table f -= a + 1; // deduct codes from patterns left xp = k; while (++j < z) { // try smaller tables up to z bits if ((f <<= 1) <= c[++xp]) { break; // enough codes to use up j bits } f -= c[xp]; // else deduct codes from patterns } } if (w + j > el && w < el) { j = el - w; // make EOB code end at table } z = 1 << j; // table entries for j-bit table lx[1 + h] = j; // set table size in stack // allocate and link in new table q = []; for (o = 0; o < z; o++) { q[o] = new HuftNode(); } if (!tail) { tail = this.root = new HuftList(); } else { tail = tail.next = new HuftList(); } tail.next = null; tail.list = q; u[h] = q; // table starts after link /* connect to last table, if there is one */ if (h > 0) { x[h] = i; // save pattern for backing up r.b = lx[h]; // bits to dump before this table r.e = 16 + j; // bits in this table r.t = q; // pointer to this table j = (i & ((1 << w) - 1)) >> (w - lx[h]); u[h - 1][j].e = r.e; u[h - 1][j].b = r.b; u[h - 1][j].n = r.n; u[h - 1][j].t = r.t; } } // set up table entry in r r.b = k - w; if (pidx >= n) { r.e = 99; // out of values--invalid code } else if (p[pidx] < s) { r.e = (p[pidx] < 256 ? 16 : 15); // 256 is end-of-block code r.n = p[pidx++]; // simple code is just the value } else { r.e = e[p[pidx] - s]; // non-simple--look up in lists r.n = d[p[pidx++] - s]; } // fill code-like entries with r // f = 1 << (k - w); for (j = i >> w; j < z; j += f) { q[j].e = r.e; q[j].b = r.b; q[j].n = r.n; q[j].t = r.t; } // backwards increment the k-bit code i for (j = 1 << (k - 1); (i & j) !== 0; j >>= 1) { i ^= j; } i ^= j; // backup over finished tables while ((i & ((1 << w) - 1)) !== x[h]) { w -= lx[h]; // don't need to update q h--; } } } /* return actual size of base table */ this.m = lx[1]; /* Return true (1) if we were given an incomplete table */ this.status = ((y !== 0 && g !== 1) ? 1 : 0); } /* routines (inflate) */ function GET_BYTE() { if (inflate_data.length === inflate_pos) { return -1; } return inflate_data[inflate_pos++] & 0xff; } function NEEDBITS(n) { while (bit_len < n) { bit_buf |= GET_BYTE() << bit_len; bit_len += 8; } } function GETBITS(n) { return bit_buf & MASK_BITS[n]; } function DUMPBITS(n) { bit_buf >>= n; bit_len -= n; } function inflate_codes(buff, off, size) { // inflate (decompress) the codes in a deflated (compressed) block. // Return an error code or zero if it all goes ok. var e; // table entry flag/number of extra bits var t; // (HuftNode) pointer to table entry var n; if (size === 0) { return 0; } // inflate the coded data n = 0; for (;;) { // do until end of block NEEDBITS(bl); t = tl.list[GETBITS(bl)]; e = t.e; while (e > 16) { if (e === 99) { return -1; } DUMPBITS(t.b); e -= 16; NEEDBITS(e); t = t.t[GETBITS(e)]; e = t.e; } DUMPBITS(t.b); if (e === 16) { // then it's a literal wp &= WSIZE - 1; buff[off + n++] = slide[wp++] = t.n; if (n === size) { return size; } continue; } // exit if end of block if (e === 15) { break; } // it's an EOB or a length // get length of block to copy NEEDBITS(e); copy_leng = t.n + GETBITS(e); DUMPBITS(e); // decode distance of block to copy NEEDBITS(bd); t = td.list[GETBITS(bd)]; e = t.e; while (e > 16) { if (e === 99) { return -1; } DUMPBITS(t.b); e -= 16; NEEDBITS(e); t = t.t[GETBITS(e)]; e = t.e; } DUMPBITS(t.b); NEEDBITS(e); copy_dist = wp - t.n - GETBITS(e); DUMPBITS(e); // do the copy while (copy_leng > 0 && n < size) { copy_leng--; copy_dist &= WSIZE - 1; wp &= WSIZE - 1; buff[off + n++] = slide[wp++] = slide[copy_dist++]; } if (n === size) { return size; } } method = -1; // done return n; } function inflate_stored(buff, off, size) { /* "decompress" an inflated type 0 (stored) block. */ var n; // go to byte boundary n = bit_len & 7; DUMPBITS(n); // get the length and its complement NEEDBITS(16); n = GETBITS(16); DUMPBITS(16); NEEDBITS(16); if (n !== ((~bit_buf) & 0xffff)) { return -1; // error in compressed data } DUMPBITS(16); // read and output the compressed data copy_leng = n; n = 0; while (copy_leng > 0 && n < size) { copy_leng--; wp &= WSIZE - 1; NEEDBITS(8); buff[off + n++] = slide[wp++] = GETBITS(8); DUMPBITS(8); } if (copy_leng === 0) { method = -1; // done } return n; } function inflate_fixed(buff, off, size) { // decompress an inflated type 1 (fixed Huffman codes) block. We should // either replace this with a custom decoder, or at least precompute the // Huffman tables. // if first time, set up tables for fixed blocks if (!fixed_tl) { var i; // temporary variable var l = []; // 288 length list for huft_build (initialized below) var h; // HuftBuild // literal table for (i = 0; i < 144; i++) { l[i] = 8; } for (null; i < 256; i++) { l[i] = 9; } for (null; i < 280; i++) { l[i] = 7; } for (null; i < 288; i++) { // make a complete, but wrong code set l[i] = 8; } fixed_bl = 7; h = new HuftBuild(l, 288, 257, cplens, cplext, fixed_bl); if (h.status !== 0) { console.error("HufBuild error: " + h.status); return -1; } fixed_tl = h.root; fixed_bl = h.m; // distance table for (i = 0; i < 30; i++) { // make an incomplete code set l[i] = 5; } fixed_bd = 5; h = new HuftBuild(l, 30, 0, cpdist, cpdext, fixed_bd); if (h.status > 1) { fixed_tl = null; console.error("HufBuild error: " + h.status); return -1; } fixed_td = h.root; fixed_bd = h.m; } tl = fixed_tl; td = fixed_td; bl = fixed_bl; bd = fixed_bd; return inflate_codes(buff, off, size); } function inflate_dynamic(buff, off, size) { // decompress an inflated type 2 (dynamic Huffman codes) block. var i; // temporary variables var j; var l; // last length var n; // number of lengths to get var t; // (HuftNode) literal/length code table var nb; // number of bit length codes var nl; // number of literal/length codes var nd; // number of distance codes var ll = []; var h; // (HuftBuild) // literal/length and distance code lengths for (i = 0; i < 286 + 30; i++) { ll[i] = 0; } // read in table lengths NEEDBITS(5); nl = 257 + GETBITS(5); // number of literal/length codes DUMPBITS(5); NEEDBITS(5); nd = 1 + GETBITS(5); // number of distance codes DUMPBITS(5); NEEDBITS(4); nb = 4 + GETBITS(4); // number of bit length codes DUMPBITS(4); if (nl > 286 || nd > 30) { return -1; // bad lengths } // read in bit-length-code lengths for (j = 0; j < nb; j++) { NEEDBITS(3); ll[border[j]] = GETBITS(3); DUMPBITS(3); } for (null; j < 19; j++) { ll[border[j]] = 0; } // build decoding table for trees--single level, 7 bit lookup bl = 7; h = new HuftBuild(ll, 19, 19, null, null, bl); if (h.status !== 0) { return -1; // incomplete code set } tl = h.root; bl = h.m; // read in literal and distance code lengths n = nl + nd; i = l = 0; while (i < n) { NEEDBITS(bl); t = tl.list[GETBITS(bl)]; j = t.b; DUMPBITS(j); j = t.n; if (j < 16) { // length of code in bits (0..15) ll[i++] = l = j; // save last length in l } else if (j === 16) { // repeat last length 3 to 6 times NEEDBITS(2); j = 3 + GETBITS(2); DUMPBITS(2); if (i + j > n) { return -1; } while (j-- > 0) { ll[i++] = l; } } else if (j === 17) { // 3 to 10 zero length codes NEEDBITS(3); j = 3 + GETBITS(3); DUMPBITS(3); if (i + j > n) { return -1; } while (j-- > 0) { ll[i++] = 0; } l = 0; } else { // j === 18: 11 to 138 zero length codes NEEDBITS(7); j = 11 + GETBITS(7); DUMPBITS(7); if (i + j > n) { return -1; } while (j-- > 0) { ll[i++] = 0; } l = 0; } } // build the decoding tables for literal/length and distance codes bl = lbits; h = new HuftBuild(ll, nl, 257, cplens, cplext, bl); if (bl === 0) { // no literals or lengths h.status = 1; } if (h.status !== 0) { if (h.status !== 1) { return -1; // incomplete code set } // **incomplete literal tree** } tl = h.root; bl = h.m; for (i = 0; i < nd; i++) { ll[i] = ll[i + nl]; } bd = dbits; h = new HuftBuild(ll, nd, 0, cpdist, cpdext, bd); td = h.root; bd = h.m; if (bd === 0 && nl > 257) { // lengths but no distances // **incomplete distance tree** return -1; } /* if (h.status === 1) { // **incomplete distance tree** } */ if (h.status !== 0) { return -1; } // decompress until an end-of-block code return inflate_codes(buff, off, size); } function inflate_start() { if (!slide) { slide = []; // new Array(2 * WSIZE); // slide.length is never called } wp = 0; bit_buf = 0; bit_len = 0; method = -1; eof = false; copy_leng = copy_dist = 0; tl = null; } function inflate_internal(buff, off, size) { // decompress an inflated entry var n, i; n = 0; while (n < size) { if (eof && method === -1) { return n; } if (copy_leng > 0) { if (method !== STORED_BLOCK) { // STATIC_TREES or DYN_TREES while (copy_leng > 0 && n < size) { copy_leng--; copy_dist &= WSIZE - 1; wp &= WSIZE - 1; buff[off + n++] = slide[wp++] = slide[copy_dist++]; } } else { while (copy_leng > 0 && n < size) { copy_leng--; wp &= WSIZE - 1; NEEDBITS(8); buff[off + n++] = slide[wp++] = GETBITS(8); DUMPBITS(8); } if (copy_leng === 0) { method = -1; // done } } if (n === size) { return n; } } if (method === -1) { if (eof) { break; } // read in last block bit NEEDBITS(1); if (GETBITS(1) !== 0) { eof = true; } DUMPBITS(1); // read in block type NEEDBITS(2); method = GETBITS(2); DUMPBITS(2); tl = null; copy_leng = 0; } switch (method) { case STORED_BLOCK: i = inflate_stored(buff, off + n, size - n); break; case STATIC_TREES: if (tl) { i = inflate_codes(buff, off + n, size - n); } else { i = inflate_fixed(buff, off + n, size - n); } break; case DYN_TREES: if (tl) { i = inflate_codes(buff, off + n, size - n); } else { i = inflate_dynamic(buff, off + n, size - n); } break; default: // error i = -1; break; } if (i === -1) { if (eof) { return 0; } return -1; } n += i; } return n; } function inflate(arr) { var buff = [], i; inflate_start(); inflate_data = arr; inflate_pos = 0; do { i = inflate_internal(buff, buff.length, 1024); } while (i > 0); inflate_data = null; // G.C. return buff; } return inflate }()); // magic numbers marking this file as GZIP var 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; } 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 = 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; } // lang ='eng' return (function(lang){ var xhr = new XMLHttpRequest(); xhr.open('GET', 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/'+lang+'.traineddata.gz', true); xhr.responseType = 'arraybuffer'; xhr.onerror = function(){ cb(xhr, null) } xhr.onload = function(){ if (xhr.status == 200 || (xhr.status == 0 && xhr.response)) { var arr = new Uint8Array(xhr.response) console.log(arr.length) window.result = new Uint8Array(unzip(arr)) console.log(result.length) } else cb(xhr, null); } xhr.send(null) }) })() setlang('eng') </script>