Decode HTML entity characters in a string into their proper character forms. Focus is on package size (~10kb gzipped) rather than speed, although it's certainly not slow.
Supported formats:
ÆæÆÆ
It does not support non-strict codes (e.g. & without ;) because this causes more problems than it solves.