Changelog

New: automatic entity escaping for &, <, and > characters. In addition, quotes " in attributes are also escaped (see #41). Prior to this, users had to provide their own filter manually. Note that jstoxml makes an effort not to escape entities that appear to have already been encoded, to prevent double-encoding issues.
- E.g. toXML({ foo: '1 < 2 & 2 > 1' }); // -> "<foo>1 < 2 & 2 > 1</foo>"
- To restore the default behavior from v1.x.x, simply pass in false to filter and attributesFilter options: toXML({ foo: '1 < 2 & 2 > 1' }, { filter: false, attributesFilter: false }); // -> "<foo>1 < 2 & 2 > 1</foo>"

fix for #40. Previously top-level objects and arrays were concatenated without proper line breaks.

Complete rewrite! The code should now be easier to understand and maintain.
now supports emoji/UTF8 tag attributes (needed for AMP pages - e.g. <html ⚡ lang="en">) (see example 14)
now supports duplicate attribute key names (see example 15)
Fixed: functions returning objects now have now that output passed through toXML for XML conversion
Fixed: empty text strings now properly output self-closing tags
Migrated tests to mocha