Proof of testnet ether usage to fight faucet drainers

[ligi]

Faucet draining is a real problem - there are people out there trying to make a profit with testnet ether.

• CAPTCHAs only hold up a little - seems people are paying to solve them. (maybe we should also try finding harder ones)
• restricting to only ENS seems to work a bit better (but comes with challenges like make clear the ENS restriction is "Mainnet ENS" #37 and is really hindering onboarding completely new users)
• inflating the total supply has not yet proven to scare faucet drainers/speculators away

But I guess even for ENS we will see draining bots over time - so I think of the current timeout-based approach might also be problematic in the future. Maybe we should think about analyzing if an ENS address really uses the testnet ether (currently the drainers are easy to detect - but might also become harder in the future)