Motivation

Truffle is a fantastic tool for creating and deploying smart contracts. We needed a way to commit deployed contract addresses as part of the repository without committing the Truffle artifacts, as they contain paths specific to each developer's filesystem.

Having the addresses separated by network allows us to ignore the local environment but commit the testnet and mainnet environments to the repository. Our continuous deployment server can then re-compile the artifacts and use the apply-registry command to merge in the deployed (Ropsten, Mainnet, etc.) addresses.

Install

``
$ npm install --save-dev truffle-deploy-registry
`

or

`
$ yarn add truffle-deploy-registry -D
`




Configuration

If you are using this library via JS (instead of via the command line) you can
configure the networks, input

| Command | Description |
| --- | --- |
| setNetworksPath(path) | Sets a new networks path (ie. 'networks-two') |
| getNetworksPath() | Returns the configured networks path (or the default networks path) |

Example of using a different path for network configs:

`javascript
var tdr = require('truffle-deploy-registry')
tdr.config.setNetworksPath('networks-two')

const networkId = 3 // ropsten
const contractName = 'SimpleToken'

// Will search the ./networks-two directory for a file called 3.json, and if there
// are multiple contract addresses listed for 'SimpleToken' it will return the most
// recent address. (Networks files are sorted chronologically)
const mostRecentAddress = findLastByContractName(networkId, contractName)

// Example return:
// {
// "contractName": "SimpleToken",
// "address": "0x8f0483125fcb9aaaefa9209d8e9d7b9c8b9fb90f",
// }
`

You can also configure the input and output artifacts path via the config object,
however those settings currently only affect the command line.




$3

After Truffle compiles your smart contracts you can merge the deployed addresses
into the artifacts by calling apply-registry from the terminal:

`sh
$ apply-registry
`

By default, apply-registry will use the truffle artifact directory './build/contracts' and the network config directory './networks'.

##### Customizing apply-registry Using Options

You can configure the input-artifacts, output-artifacts and networks directories
which apply-registry uses via the command line options. For example:

`sh
$ apply-registry -i build/contracts -o build/output -n networks
`

In this case, merged artifacts would appear in build/output instead of build/contracts.




Usage

Truffle Deploy Registry works in two stages:

1. New deployment entries are recorded in a network-specific JSON file.
2. The latest deployment entries are merged with the truffle artifacts after compilation.

$3

Truffle Deploy Registry stores contract addresses in JSON files in the networks/ directory. For example, if you deploy to mainnet and ropsten your networks directory may look like:

`
networks/
1.json
3.json
`

Each of these files contains an array of deployment entries. New entries are appended. Each entry must store the contractName and address, but is otherwise unstructured so that the user can add additional information. A network config looks something like:

`


networks/1.json


[
{ contractName: 'Ownable', address: '0x3383c29542b8c96eafed98c4aafe789ddb256e19', transactionHash: '0x0b71a01c6da8e02359b533f16b97a590be8ca59480151ba1034a264a2981261f' },
{ contractName: 'Registry', address: '0x8fa5944b15c1ab5db6bcfb0c888bdc6b242f0fa6', transactionHash: '0x84a9fe87a9fd8f7ae98fa72359b533f16b97a590be8ca59480151ba1034a2632' }
]
`

To add new entries call the appendInstance function:

`javascript
// migrations/1_initial_migration.js

var tdr = require('truffle-deploy-registry')
var Migrations = artifacts.require("./Migrations.sol");

module.exports = function(deployer, network) {
deployer.deploy(Migrations).then((migrationsInstance) => {
if (!tdr.isDryRunNetworkName(network)) {
return tdr.appendInstance(migrationsInstance)
}
})
}
`

Alternatively, you can use the lower-level append function:

`javascript
// migrations/1_initial_migration.js

var tdr = require('truffle-deploy-registry')
var Migrations = artifacts.require("./Migrations.sol");

module.exports = function(deployer, network) {
deployer.deploy(Migrations).then(async (migrationsInstance) => {
if (!tdr.isDryRunNetworkName(network)) {
await tdr.append(deployer.network_id, {
contractName: 'Migrations',
address: Migrations.address,
transactionHash: migrationsInstance.transactionHash
})
}
})
}
`

Note the use of isDryRunNetworkName to prevent appending to the registry when doing a dry run.

$3

After Truffle compiles your smart contracts, you can merge the deployed addresses into the artifacts by calling apply-registry from the terminal:

`sh
$ apply-registry
`

NOTE: By default, apply-registry will use the truffle artifact directory './build/contracts' and the network config directory './networks'. If you need to customize this see apply-registry options

This will pull in all of the network configs and add the most recent address for each contract by name from each configuration. For example, if you have two configs:

`
networks/
1.json
3.json
`

1.json:

`json
[
{ "contractName": "Contract2", "address": "0x2222222222222222222222222222222222222222", "transactionHash": "0x0b71a01c6da8e02359b533f16b97a590be8ca59480151ba1034a264a2981261f" },
{ "contractName": "Contract2", "address": "0x4444444444444444444444444444444444444444", "transactionHash": "0x21afe897aefa98eaf79ae6f87ae6f87678e6f39480151ba1034a264a29853124" },
]
`

3.json:

`json
[
{ "contractName": "Contract2", "address": "0x8888888888888888888888888888888888888888", "transactionHash": "0x99afe897aefa98eaf79ae6f87ae6f87678e6f39480151ba1034a264a29853124" }
]
`

build/contracts/Contract2.json:

`json
{
"contractName": "Contract2",
"abi": [],
"networks": {
"2": {
"events": {},
"links": {},
"address": "0x3383c29542b8c96eafed98c4aafe789ddb256e19",
"transactionHash": "0xbef1787981c4512c8bc8acf59e6cbe9c23c9c5ea269b193ec71aae9f9c57c997"
}
},
"updatedAt": "2018-09-19T21:37:52.903Z"
}
`

Then run apply-registry:

`sh
$ apply-registry
`

Your artifact build/contracts/Contract2.json will now be updated with the networks:

`json
{
"contractName": "Contract2",
"abi": [],
"networks": {
"1": {
"events": {},
"links": {},
"address": "0x4444444444444444444444444444444444444444",
"transactionHash": "0x21afe897aefa98eaf79ae6f87ae6f87678e6f39480151ba1034a264a29853124"
},
"2": {
"events": {},
"links": {},
"address": "0x3383c29542b8c96eafed98c4aafe789ddb256e19",
"transactionHash": "0xbef1787981c4512c8bc8acf59e6cbe9c23c9c5ea269b193ec71aae9f9c57c997"
},
"3": {
"events": {},
"links": {},
"address": "0x8888888888888888888888888888888888888888",
"transactionHash": "0x99afe897aefa98eaf79ae6f87ae6f87678e6f39480151ba1034a264a29853124"
},
},
"updatedAt": "2018-09-19T21:37:52.903Z"
}
`

It can help to create a new script entry in package.json so that you can easily combine compilation with merging:

`json
{
"scripts": {
"compile": "truffle compile && apply-registry"
}
}
`

##### Another Use Case

Some of our contract projects use ZeppelinOS. While ZOS has many upsides (easily upgradeable contracts being a huge one, guards against destroying deployed memory addresses, etc) we've found the need to manually copy over contract addresses both in development and on testnets/mainnet.

The apply-registry tool will generate new artifacts in the output artifacts directory even if there are no input artifacts. The workflow for this is as follows:

1. Deploy contracts using ZOS to local ganache, ropsten, etc.

2. Manually copy the deployed contract names and addresses to a new networks file in our dapp directory. For instance, the networks file: ./networks/3.json would have entries as such (where the lowest entry address '0xae399886...' is the most recent deployed version):

`javascript
[
{
"contractName": "SimpleToken",
"address": "0x8f0483125fcb9aaaefa9209d8e9d7b9c8b9fb90f"
},
{
"contractName": "SimpleToken",
"address": "0xae39986e9876c91a936adfae8b6a98e764aeeb7a"
}
]
`

3. Run apply-registry and it will generate the SimpleToken.json artifact in './build/contracts', ready for the dapp to use.
4. To get the ABI, use npm or yarn's link command. Or, if the contracts are published on npm simply add the contracts package to your package.json and install.

$3

You can retrieve the last entry by contract name using the findLastByContractName(networkId, contractName) function.

For example:

`javascript
const TestContract = artifacts.require('TestContract.sol')
const tdr = require('truffle-deploy-registry')

module.exports = function(deployer, networkName) {
deployer.then(async () => {
const entry = await tdr.findLastByContractName(deployer.network_id, 'Contract2')
deployer.deploy(TestContract, entry.address).then(instance => {
if (!tdr.isDryRunNetworkName(networkName)) {
await tdr.appendInstance(instance)
}
})
})
}
``

Future Work

Eventually it would be best to create entirely separate artifacts that include the bytecode and abi, rather than having to merge into the compiled artifact. Another possibility is to integrate with EthPM and treat each deployment as a package.