Support for resolving TLDs with punycodes

[apbigcod]

The current implementation is not able to resolve root domains with puny codes.

ENS supports unicode characters and therefore TLDs with special characters can be directly registered as they are without mapping to punycodes or other encodings.

As an example, if we take ☢㣎, punycode xn--j4h059d, we can create an ENS registry with ☢㣎 whose hash is 0x9944186df8dbb5bd89fd1d89d38b3889b75f7aa9fd3a0b66921a49372935804c.

When typing subdomain.☢㣎 in a browser, it will be mapped to subdomain.xn--j4h059d which will trigger fingertip to query the handshake record for xn--j4h059d resulting in a contract address which then needs to be queries using ☢㣎 rather than its punycode equivalent.

In short the flow currently is:

1. type subdomain.☢㣎 in brower
2. browser maps to domain.xn--j4h059d
3. Fingertip queries handshake for .xn--j4h059d and gets ENS registry contract address
4. Fingertip queries contract for domain.xn--j4h059d

But step 4 should be:
4. Fingertip queries contract for domain.☢㣎

This PR normalises the name to resolve before we call the ENS contract.

[buffrr]

LGTM thanks!