Fix for test failure: encoding to UTF-16 is implementation dependent,…

… output may be either big or little endian. Therefore, encode explicitly to UTF-16LE.

test-big-buffer.js

@@ -1,23 +1,18 @@
 var Iconv = require('iconv').Iconv,
    assert = require('assert');
 
-var iconv = new Iconv('UTF-8', 'UTF-16');
+var iconv = new Iconv('UTF-8', 'UTF-16LE');
 
 var utf8 = new Buffer(20000000);
 for (var i = 0; i < utf8.length; i++) {
   utf8[i] = 97 + i % 26; // cycle from 'a' to 'z'.
 }
 var utf16 = iconv.convert(utf8);
 
-// 2 bytes per character + 2 bytes for BOM
-assert.equal(utf16.length, utf8.length * 2 + 2);
-
-// test for BOM
-assert.equal(utf16[0], 0xff);
-assert.equal(utf16[1], 0xfe);
+assert.equal(utf16.length, utf8.length * 2);
 
 for (i = 0; i < utf8.length; ++i) {
-  assert.equal(utf16[2 + i * 2], utf8[i]);
-  assert.equal(utf16[2 + i * 2 + 1], 0);
+  assert.equal(utf16[i * 2], utf8[i]);
+  assert.equal(utf16[i * 2 + 1], 0);
 }