JSON prettifier breaks HTML encoded entities for non-JSON

[alex-sherwin]

The way the current JSON prettifier logic works, it reads the contents as text, checks if it's JSON, if it is it prettifies it, then it always sets the data back.

It shouldn't be resetting the data it if didn't modify it (i.e. if it's not JSON), because this process will lose HTML entity encoding (most apparent when rendering XML responses, as it won't work at all as-is, since the < and > characters lose their encoding)