Fix hmt encoding

[ShellyHerself]

Reported issue:

    Could not load hmt file.
Creating hud_message_text from this hmt file:
    C:\Users\kianh\OneDrive\Documents\camo_function.hmt
    WARNING: Message name on line 1 is too long. Truncating name to '扯彪慣潭敲档牡楧杮㴠䌠浡敲档牡楧杮⸮മ漊橢损浡牯慥祤㴠䌠浡'
    ERROR: No name or message on line 1

[ShellyHerself]

Going to update this a little more

[ShellyHerself]

Okay. I am confident of this now. The detection library that I have put in place for this is known to detect 97.25% of human readable files correctly. The only place I've seen it struggle is code files that are specifically utf-16.

So, as a fallback, whenever it can't detect what encoding something is, it picks utf-16 as that is what the expected format for these files was in the first place.

[ShellyHerself]

I am not pulling this library in for just this btw. There is some other stuff with string encoding that I want to improve later on. Like how the string list importer creates either a unicode stringlist or ascii stringlist based on the guessed file encoding while these two tag types are not at all interchangable.

I will later on be separating those two functions in Mozzarilla so that people can pick exactly what they need.

Lower the bar of entry you could say.