Add ReverseNamer

[adraffy]

• added ReverseNamer.sol and tests
  1. NamedOnce for simple immutable name on constructor
  2. NameableBy for delegated naming
  3. ReverseNamer.setName("<name>") for custom implementation

Usage:

// 1
import {NamedOnce} from "@ensdomains/ens-contracts/contracts/reverseRegistrar/ReverseNamer.sol";
contract C is NamedOnce("mycontract.eth") { ... }

// 2
import {NameableBy} from "@ensdomains/ens-contracts/contracts/reverseRegistrar/ReverseNamer.sol";
contract C is NameableBy {
    constructor(address namer) NameableBy(namer, "") { ... }
}

// 3
import {ReverseNamer} from "@ensdomains/ens-contracts/contracts/reverseRegistrar/ReverseNamer.sol";
contract C {
    function f() external {
        ReverseNamer.setName("mycontract.eth");
    }
}

---

Alternative ideas:

• NamedOnce → emit Named(string) → generate proof of log
• code snippet: address(<registrar>).call(abi.encodeWithSignature("setName(string)", "<name>"))

[gskril]

I like it! Should this PR also delete ReverseClaimer.sol, or maybe more safely just add a natspec comment indicating to use ReverseNamer.sol for wider support (rollups)?

[NickSneo]

@adraffy Whats the difference between case 2 and case 3
isn't it indirectly doing same thing delegating someone to name the contract either by directly delegating in case 2 or indirectly (via some access control on who can call function f()) in case 3

Would it makes sense to just have 2 cases?

[NickSneo]

PR looks good, thanks for this. We as Enscribe team also need to push more on awareness so developers include this in their contracts before deploying.

Added a few comments

[NickSneo]

need to add TESTNET or is it same as MAINNET address? since we have distinct MAINNET_ROLLUP and TESTNET_ROLLUP

[adraffy]

I think this will mean all mainnets (mainnet/sepolia/...).

I speculatively designed ETHReverseResolver as the v2 replacement for addr.reverse that uses the same tech as the rollups (independent registry + wildcard resolver) but I think we're going with a proper claimable {addr}.addr.reverse that uses the v2 design, but it should be a fixed address deployment and the same on those chains. And If not, you're correct.

[NickSneo]

on line 10 - MAINNET = address(0) so is the else case

[NickSneo]

NotOwnable in case of NamedOnce makes sense as no one can later name the contract but in case of NameableBy doesnt make much sense to it.

NameableBy can be seen in a contract with Ownable as an extra delegate who can name contract but still if contract has Owner he can revoke the delegation. I might be wrong but limiting NameableBy to only non ownable contracts doesn't make much sense to me

[adraffy]

Yeah, I wasn't sure about this design. NameableBy could be removed. My thinking was you want something like Ownable but without implementing Ownable which at a glance implies the contract is mutable.

NameableBy + Ownable would give you 2 accounts that can change the name, but maybe that should be allowed.

I was thinking of it more like a trust curve:

1. NamedOnce — set once
2. NameableBy — set later, revoke it / delegate to someone to name it
3. NameableBy — delegate to someone, change whenever
4. Ownable — change whenever, but shares with other Ownable responsibilities