doesn't decode &Amp; - purposeful? about htmlentities

[threedaymonk]

Unfortunately, there's not a simple solution to this, because the case of entity names matters: Ü (Ü) is different from ü (ü).

I don't know that there is a good solution. Writing something that falls back to a case-insensitive match when the case-sensitive match fails would be possible, but I doubt it has enough general utility to be worthwhile: this is the first time I've seen this problem. Some kind of custom mapping is probably the way to go, but whilst I've played around with some thoughts in that direction, the API doesn't facilitate that at present.

In your situation, I think I might tack on an extra post-processing step that catches and fixes any errant &Amp; along with anything else you come across.