What's the difference between upgradable and non upgradable Initializable contracts?

Question

@openzeppelin/contracts/proxy/utils/Initializable.sol

vs

@openzeppelin/contracts-upgradeable/proxy/utils/Initializable.sol

Why would anyone use proxies that do not support upgradable contracts? Instead, just create contract without a proxy, doesn't that save the trouble?

// SPDX-License-Identifier: MIT
pragma solidity 0.8.15;

import { Initializable } from "@openzeppelin/contracts/proxy/utils/Initializable.sol";

contract Test is Initializable {
    address public owner;

    constructor() public {
        owner = msg.sender;
    }
    function initialize()
        public
        initializer
    {
    // logic
    }
}

Aren't all proxies meant to support upgradable implementation contracts??

Another question is since it uses a non-upgradable contract, is it still susceptible to the constructor caveat? Openzeppelin Docs

Answer 1

Proxies can be implemented with and without the upgradeability feature (see Proxies). Depending on the use case or technical considerations, you may opt to deploy a non-upgradeable proxy using one of the most known mechanism such as Diamond Proxy (see ERC-2535).

One use case of a non-upgradeable proxy would be to separate your contract's functions into smaller contracts called facets and allow your proxy to route functions call to the proper facet. By doing so, for example, you can get rid of the maximum contract size limit (which is now 24.576 kb).

Answer 2

1. @openzeppelin/contracts/proxy/utils/Initializable.sol and @openzeppelin/contracts-upgradeable/proxy/utils/Initializable.sol is the same, diffrent import from diffrent package

2. When using upgradable we don't use the constructor to prevent re-initialize contract state

  /// @custom:oz-upgrades-unsafe-allow constructor
  constructor() {
    _disableInitializers();
  }