All node response parsers assume utf8 encoding

I noticed [all parsers assume utf8](https://github.com/visionmedia/superagent/blob/e11119b0b6956e3a57d24c9093ee3057d33bde62/lib/node/parsers/text.js#L4).

e.g. text parser:

``` javascript
module.exports = function(res, fn){
res.text = '';
res.setEncoding('utf8');
res.on('data', function(chunk){ res.text += chunk; });
res.on('end', fn);
};
```

It seems like they all should get the encoding from the response content-type header and fall back to binary (application/octet-stream) if none exists.

Can PR, if desired.