XMLStringifier does not handle HTML character entities

[feedler]

The stringifier replaces all occurences of & with & but does not consider that the string might contain HTML character entities itself (like for example ü or ü for the umlaut ü). The following change in the (javascript-) source code should fix this:

// Current source:
return str.replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/\r/g, '&#xD;');
// Fixed source:
return str.replace(/(?!&(\D+|#\d+);)&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/\r/g, '&#xD;');

[feedler]

Had a mistake in the regex myself. This version should work now:

return str.replace(/(?!&#?\S+;)&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/\r/g, '&#xD;');

[oozcitak]

Are you aiming to avoid double encoding of &? Also isn't /(?!&\S+;)&/g sufficient in your regex?

[feedler]

Yes - that's what I'm aiming at. And you're right: your regex should be sufficient. :)

[oozcitak]

I am going to add a doubleEncode flag for this. It will default to true so that current default behavior will remain unchanged. When explicitly set to false it will use your regex, Would that work for you?

[feedler]

Yes, that will work - thank you very much!

[oozcitak]

You can now create the builder with noDoubleEncoding: true to preserve HTML entities.
See: https://github.com/oozcitak/xmlbuilder-js/blob/master/test/issues/97.coffee#L9