The working group came to the same conclusion. Whatever crap the client puts in the packet is inviolate. Don't change it. You're not smart enough to do that.
Instead, if you need to do comparisons, compare on the normalized strings. You create a "normalized" version of the string, and then use that for comparisons.
I wish the IETF precis working group had a "recommendations" document. It would make everyone's lives much simpler.
I spent the last 5 years trying to get i18n working in a protocol:
https://tools.ietf.org/html/draft-ietf-radext-nai-15
The working group came to the same conclusion. Whatever crap the client puts in the packet is inviolate. Don't change it. You're not smart enough to do that.
Instead, if you need to do comparisons, compare on the normalized strings. You create a "normalized" version of the string, and then use that for comparisons.
I wish the IETF precis working group had a "recommendations" document. It would make everyone's lives much simpler.