Guides
Extending World

Extending World

On this page you learn how to modify a World that is already deployed to the blockchain. If you want to learn how to modify a World before it is deployed, see the hello world page.

The sample program

To learn how to extend a World, we will extend the Counter example to allow users to leave a message while incrementing the counter. The steps to create the example World that we'll extend are here. You can either do it now, or wait until after you've created the extension resources.

To extend the World with the message functionality requires adding several resources:

  • Namespace. A namescape can contain tables and systems. In most cases, the only way you would be able to extend a World that somebody else owns is to create your own namespace within that world.

  • Table. A table to store the messages that have been sent.

  • System. A system that updates the messages table and then calls increment to update the counter.

Create the Solidity code

The easiest way to create the Solidity code is to use the MUD template:

  1. Create a new MUD application. Use either the vanilla template or the react-ecs one.

    pnpm create mud@next extension
    cd extension/packages/contracts
  2. Edit mud.config.ts to include the definitions we need.

    mud.config.ts
    import { mudConfig } from "@latticexyz/world/register";
     
    export default mudConfig({
      namespace: "messaging",
      tables: {
        Messages: {
          keySchema: {
            counterValue: "uint32",
          },
          valueSchema: {
            message: "string",
          },
        },
      },
      systems: {
        MessageSystem: {
          name: "MessageSystem",
          openAccess: true,
        },
      },
    });
  3. Create src/systems/MessageSystem.sol.

    MessageSystem.sol
    // SPDX-License-Identifier: MIT
    pragma solidity >=0.8.21;
     
    import { System } from "@latticexyz/world/src/System.sol";
    import { Messages } from "../codegen/index.sol";
     
    interface WorldWithIncrement {
      function increment() external returns (uint32);
    }
     
    contract MessageSystem is System {
      function incrementMessage(string memory message) public returns (uint32) {
        uint32 newVal = WorldWithIncrement(_world()).increment();
        Messages.set(newVal, message);
        return newVal;
      }
    }
    Explanation
    import { System } from "@latticexyz/world/src/System.sol";
    import { Messages } from "../codegen/index.sol";

    These are the two main things the System needs to know: how to be a System and how to access the Messages table.

    interface WorldWithIncrement {
      function increment() external returns (uint32);
    }

    This System needs to call increment on the World where it is implemented. However, as an extension author you might not have access to the source code of any System that isn't part of your extension.

    If you define your own interface for World you can add whatever function signatures are supported. Note that in Ethereum a function signature (opens in a new tab) is the function name and its parameter types, it does not include the return type. So if you are unsure of the return type that is not a huge problem.

    contract MessageSystem is System {
      function incrementMessage(string memory message) public returns (uint32) {
        uint32 newVal = WorldWithIncrement(_world()).increment();

    This is how we use the WorldWithIncrement interface we created. The _world() (opens in a new tab) call gives us the address of the World that called us. When we specify WorldWithIncrement(<address>), we are telling Solidity that there is already a WorldWithIncrement at that address, and therefore we can use functions that are supported by WorldWithIncrement, such as increment().

        Messages.set(newVal, message);

    This is one way to create a record with the key newVal and the value message.

        return newVal;
      }
    }

    When we are called by a user, an externally owned account, the return value is meaningless. However, just as we call increment() from a contract and use the return value (instead of having to read Counter ourselves), some future onchain code might call incrementMessage and use the returned value.

  4. Remove files that are part of the template but don't make sense when our extension doesn't have its own counter.

    rm src/systems/IncrementSystem.sol test/*.t.sol script/PostDeploy.s.sol
  5. Build and compile the Solidity code.

    pnpm build

Deploy to the blockchain

  1. To have a Counter example World to modify, go to a separate command line window and create a blockchain with a World using the TypeScript template and start the execution. Choose either the react user interface or the vanilla one.

    pnpm create mud@next extendMe
    cd extendMe
    pnpm dev
  2. Back in the extension's packages/contracts directory, create a .env file with:

    • PRIVATE_KEY - the private key of an account that has ETH on the blockchain.
    • WORLD_ADDRESS - the address of the World to which you add the namespace.

    If you are using the template with a fresh pnpm dev, then you can use this .env:

    .env
    # Anvil default private key for the second account
    # (NOT the account that deployed the World)
    PRIVATE_KEY=0x59c6995e998f97a5a0044966f0945389dc9e86dae88c7a8412f4603b6b78690d
     
    # Address for the world we are extending
    WORLD_ADDRESS=0x6e9474e9c83676b9a71133ff96db43e7aa0a4342
  3. Create this script in script/MessagingExtension.s.sol.

    MessagingExtension.s.sol
    // SPDX-License-Identifier: MIT
    pragma solidity >=0.8.21;
     
    import { Script } from "forge-std/Script.sol";
    import { console } from "forge-std/console.sol";
    import { IBaseWorld } from "@latticexyz/world-modules/src/interfaces/IBaseWorld.sol";
     
    import { WorldRegistrationSystem } from "@latticexyz/world/src/modules/core/implementations/WorldRegistrationSystem.sol";
     
    // Create resource identifiers (for the namespace and system)
    import { ResourceId } from "@latticexyz/store/src/ResourceId.sol";
    import { WorldResourceIdLib } from "@latticexyz/world/src/WorldResourceId.sol";
    import { RESOURCE_SYSTEM } from "@latticexyz/world/src/worldResourceTypes.sol";
     
    // For registering the table
    import { Messages, MessagesTableId } from "../src/codegen/index.sol";
    import { IStore } from "@latticexyz/store/src/IStore.sol";
    import { StoreSwitch } from "@latticexyz/store/src/StoreSwitch.sol";
     
    // For deploying MessageSystem
    import { MessageSystem } from "../src/systems/MessageSystem.sol";
     
    contract MessagingExtension is Script {
      function run() external {
        uint256 deployerPrivateKey = vm.envUint("PRIVATE_KEY");
        address worldAddress = vm.envAddress("WORLD_ADDRESS");
     
        WorldRegistrationSystem world = WorldRegistrationSystem(worldAddress);
        ResourceId namespaceResource = WorldResourceIdLib.encodeNamespace(bytes14("messaging"));
        ResourceId systemResource = WorldResourceIdLib.encode(RESOURCE_SYSTEM, "messaging", "message");
     
        vm.startBroadcast(deployerPrivateKey);
        world.registerNamespace(namespaceResource);
     
        StoreSwitch.setStoreAddress(worldAddress);
        Messages.register();
     
        MessageSystem messageSystem = new MessageSystem();
        console.log("MessageSystem address: ", address(messageSystem));
     
        world.registerSystem(systemResource, messageSystem, true);
        world.registerFunctionSelector(systemResource, "incrementMessage(string)");
     
        vm.stopBroadcast();
      }
    }
    Explanation
    // SPDX-License-Identifier: MIT
    pragma solidity >=0.8.21;

    Standard Solidity boilerplate.

    import { Script } from "forge-std/Script.sol";
    import { console } from "forge-std/console.sol";

    The definitions for forge scripts (opens in a new tab) and the console (opens in a new tab).

    import { IBaseWorld } from "@latticexyz/world-modules/src/interfaces/IBaseWorld.sol";

    Use IBaseWorld.sol (opens in a new tab) to get definitions that are common to all World contracts.

    import { WorldRegistrationSystem } from "@latticexyz/world/src/modules/core/implementations/WorldRegistrationSystem.sol";

    WorldRegistartionSystem.sol (opens in a new tab) contains the function definitions necessary to register new namespaces and systems with an existing World.

    // Create resource identifiers (for the namespace and system)
    import { ResourceId } from "@latticexyz/store/src/ResourceId.sol";
    import { WorldResourceIdLib } from "@latticexyz/world/src/WorldResourceId.sol";
    import { RESOURCE_SYSTEM } from "@latticexyz/world/src/worldResourceTypes.sol";

    These definitions make it easy to manage resource identifiers. We need them for the resource IDs we need to create: the namespace and the system.

    // For registering the table
    import { Messages, MessagesTableId } from "../src/codegen/index.sol";
    import { IStore } from "@latticexyz/store/src/IStore.sol";
    import { StoreSwitch } from "@latticexyz/store/src/StoreSwitch.sol";

    These are the definitions we need to register the Messages table.

    // For deploying MessageSystem
    import { MessageSystem } from "../src/systems/MessageSystem.sol";

    These are the definitions we need to deploy the MessageSystem contract so we'll then be able to register it as a System in the World.

    contract MessagingExtension is Script {
        function run() external {

    This is the function that implements the script.

            uint256 deployerPrivateKey = vm.envUint("PRIVATE_KEY");
            address worldAddress = vm.envAddress("WORLD_ADDRESS");

    Read the private key and the address of the World from the environment (which includes the content of the .env file).

            WorldRegistrationSystem world = WorldRegistrationSystem(worldAddress);
            ResourceId namespaceResource = WorldResourceIdLib.encodeNamespace(bytes14("messaging"));
            ResourceId systemResource = WorldResourceIdLib.encode(RESOURCE_SYSTEM, "messaging", "message");

    Among other things, a MUD World is a WorldRegistrationSystem, so it has the appropriate functions. A ResourceId is a 32 byte value that uniquely identifies a resource in a MUD World. It is two bytes of resource type followed by 14 bytes of namespace and then 16 bytes of the name of the actual resource.

    Here we create two ResourceId values:

    NameTypeNamespaceResource name
    namespaceResourcens (namespace)messagingEmpty
    systemResourcesy (system)messagingmessage

    If you want to see these values, add these two lines to the script:

            console.log("Namespace ID: %x", uint256(ResourceId.unwrap(namespaceResource)));
            console.log("System ID:    %x", uint256(ResourceId.unwrap(systemResource)));

    Note that console.log requires a uint256 value, and we can't get that directly from a ResourceId. Instead, we have to unwrap (opens in a new tab) our ResourceId to get the original type (bytes32) and then cast (opens in a new tab) it to uint256.

    The expected values are:

    NameExpected value
    Namespace ID0x6e736d6573736167696e67000000000000000000000000000000000000000000
    System ID0x73796d6573736167696e6700000000006d657373616765000000000000000000

    You can use an online calculator (opens in a new tab) to verify the values are correct.

    Hex valueASCII
    6e736d6573736167696e67nsmessaging
    73796d6573736167696e67symessaging
    6d657373616765message
            vm.startBroadcast(deployerPrivateKey);

    Use the private key to submit transactions.

            world.registerNamespace(namespaceResource);

    Register the namespace (opens in a new tab).

            StoreSwitch.setStoreAddress(worldAddress);
            Messages.register();

    Register the Messages table to worldAddress.

            MessageSystem messageSystem = new MessageSystem();
            world.registerSystem(systemResource, messageSystem, true);

    Deploy the new system and then register it with the World we are extending. The last parameter is whether or not we allow everybody to access this System.

            world.registerFunctionSelector(systemResource, "incrementMessage(string)");

    Register MessageSystem.incrementMessage(string). This step is necessary to make the function accessible through the World. The function's name when accessed through the world is <namespace>_<system>_<function>, so this function will be available as messaging_message_incrementMessage(string).

    Note: This is the case for all namespaces except for the root namespace, where we just use the name of the function (such as increment()).

            vm.stopBroadcast();
        }
    }

    Stop using the private key. Here this call is not necessary because we immediately leave the script, but it is a good idea to include it in case MessagingExtension.run() ever becomes part of a larger script.

  4. Run the script. Note that you need to provide the URL in the command line, you can't rely on the ETH_RPC_URL environment variable.

    forge script script/MessagingExtension.s.sol --rpc-url http://localhost:8545 --broadcast
  5. Increment and write a message.

    source .env
    cast send $WORLD_ADDRESS --private-key $PRIVATE_KEY "messaging_message_incrementMessage(string)" "hello"

    When a function is not in the root namespace, it is accessible as <namespace>_<system>_<function name> (as long as it is registered (opens in a new tab)).

    Here we call our incrementMessage(string) with the parameter hello.

  6. You can see in the user interface of extendMe that the counter has been incremented. To see the message we sent, use these commands:

    TABLE_ID=0x74626d6573736167696e6700000000004d657373616765730000000000000000
    KEY=[`cast to-int256 2`]
    RAW_RESULT=`cast call $WORLD_ADDRESS "getRecord(bytes32,bytes32[])" $TABLE_ID $KEY`
    cast --to-ascii ${RAW_RESULT:322:-2}
    Explanation
    TABLE_ID=0x74626d6573736167696e6700000000004d657373616765730000000000000000

    TABLE_ID is the ResourceId for the table, taken from src/codegen/tables/Messages.sol You can verify the interpretation with the online calculator (opens in a new tab).

    TypeNamespaceResource name
    tb (table)messagingMessages
    KEY=[`cast to-int256 2`]

    The key to a MUD table is always an array of byte32 values. To create that value, we convert our key (2, because that's the first user call to increment(), the 0->1 increment is done by the post deploy script) to a 256 bit value using cast (opens in a new tab) and then envelop it in an array.

    RAW_RESULT=`cast call $WORLD_ADDRESS "getRecord(bytes32,bytes32[],bytes32)" $TABLE_ID $KEY $FIELD_LAYOUT`

    To call getRecord (opens in a new tab) we need the tableID, the key array, and the field layout. The call result is the entire value, which may have multiple static (fixed length) and dynamic (variable length) fields.

    Here is the raw result divided into 32 byte words.

    WordValue
    00000000000000000000000000000000000000000000000000000000000000060
    10000000000000000000000000000000000000000000000000500000000000005
    20000000000000000000000000000000000000000000000000000000000000080
    30000000000000000000000000000000000000000000000000000000000000000
    40000000000000000000000000000000000000000000000000000000000000005
    568656c6c6f000000000000000000000000000000000000000000000000000000

    When there is a variable-length field (a field whose length is not known at compile-time) in a Solidity function's return value, it is represented by a word that tells us at what offset into the return data that field starts. This function is used for different tables with different lengths of static fields, so the static fields are a variable length field as far as Solidity is concerned.

    Word 0 shows us that this field's value starts at 0x60, which is word 3. Because there are no static fields, word 3 is all zeros.

    Word 1 is a PackedCounter (opens in a new tab) with the lengths of the dynamic fields. Here is the interpretation.

    BytesValueMeaning
    6-000000000000005The total length of all dynamic fields is five bytes.
    11-70000000005The length of the first dynamic field is five bytes.
    16-120000000000The length of the (non existent in Messages) second dynamic field is zero
    21-170000000000The length of the (non existent in Messages) third dynamic field is zero
    26-220000000000The length of the (non existent in Messages) fourth dynamic field is zero
    31-270000000000The length of the (non existent in Messages) fifth dynamic field is zero

    MUD tables can only have up to five dynamic fields because PackedCounter needs to fit in a 32 byte word.

    Word 2 shows us that the field with the dynamic lengths starts at byte 0x80, which is word 4. Word 4 gives us the length of the string.

    Finally, word 5 gives us the actual message, "hello".

    cast --to-ascii ${RAW_RESULT:322:-2}

    Based on the above we don't need the first 322 characters of $RAW_RESULT. The first two characters are the 0x that tells us this is a hexadecimal number. The next 320 characters are words 0-4, which are not the actual message (each word is 32 bytes, which is 64 hexadecimal digits). We also don't need the trailing newline. ${RAW_RESULT:322:-2} removes those characters so we can use cast (opens in a new tab) to get the ASCII.