This is a description of the Flow standard for fungible token contracts. It is meant to contain the minimum requirements to implement a safe, secure, easy to understand, and easy to use fungible token contract. It also includes an example implementation to show how a concrete smart contract would actually implement the interface.
The version of the contracts in the master
branch is the
Cadence 1.0 version of the contracts and is not the same
as the ones that are currently deployed to testnet and mainnet.
See the cadence-0.42
branch for the currently deployed versions.
Flow is a new blockchain for open worlds. Read more about it here.
Cadence is a new Resource-oriented programming language for developing smart contracts for the Flow Blockchain. Read more about it here and see its implementation here
We recommend that anyone who is reading this should have already completed the Cadence Tutorials so they can build a basic understanding of the programming language.
Resource-oriented programming, and by extension Cadence, is the perfect programming environment for currencies, because users are able to store their tokens directly in their accounts and transact peer-to-peer. Please see the blog post about resources to understand why they are perfect for digital assets.
The FungibleToken
, FungibleTokenMetadataViews
, and FungibleTokenSwitchboard
contracts are already deployed
on various networks. You can import them in your contracts from these addresses.
There is no need to deploy them yourself.
Network | Contract Address |
---|---|
Emulator | 0xee82856bf20e2aa6 |
Testnet | 0x9a0766d93b6608b7 |
Sandboxnet | 0xe20612a0776ca4bf |
Mainnet | 0xf233dcee88fe0abe |
The Burner
contract is also deployed to these addresses, but should only be used in Cadence 1.0 FungibleToken
implementations of the standard.
The code for the standard is in contracts/FungibleToken.cdc
. An example implementation of the standard that simulates what a simple token would be like is in contracts/ExampleToken.cdc
.
The exact smart contract that is used for the official Flow Network Token (FlowToken
) is in the flow-core-contracts
repository.
Example transactions that users could use to interact with fungible tokens are located in the transactions/
directory. These templates are mostly generic and can be used with any fungible token implementation by providing the correct addresses, names, and values.
The standard consists of a contract interface called FungibleToken
that defines important
functionality for token implementations. Contracts are expected to define a resource
that implement the FungibleToken.Vault
resource interface.
A Vault
represents the tokens that an account owns. Each account that owns tokens
will have a Vault
stored in its account storage.
Users call functions on each other's Vault
s to send and receive tokens.
The standard uses unsigned 64-bit fixed point numbers UFix64
as the type to represent token balance information. This type has 8 decimal places and cannot represent negative numbers.
Specifies that the implementing type must have a UFix64
balance
field.
access(all) var balance: UFix64
Defines a withdraw
function for withdrawing a specific amount of tokens as amount.
access(all) fun withdraw(amount: UFix64): @{FungibleToken.Vault}
- Conditions
- the returned Vault's balance must equal the amount withdrawn
- The amount withdrawn must be less than or equal to the balance
- The resulting balance must equal the initial balance - amount
- Conditions
- Users can give other accounts a persistent Capability
or ephemeral reference to their
Vault
cast as aProvider
to allow them to withdraw and send tokens for them. A contract can define any custom logic to govern the amount of tokens that can be withdrawn at a time with aProvider
. This can mimic theapprove
,transferFrom
functionality of ERC20. FungibleToken.Withdrawn
event- Event that is emitted automatically to indicate how much was withdrawn
and from what account the
Vault
is stored in. If theVault
is not in account storage when the event is emitted,from
will benil
. - Contracts do not have to emit their own events,
the standard events will automatically be emitted from the interface contract with values identifying the relevant
Vault
.
- Event that is emitted automatically to indicate how much was withdrawn
and from what account the
Defines an isAvailableToWithdraw()
function
to ask a Provider
if the specified number of tokens can be withdrawn from the implementing type.
Defines functionality to depositing fungible tokens into a resource object.
deposit()
function:access(all) fun deposit(from: @{FungibleToken.Vault})
- Conditions
from
balance must be non-zero- The resulting balance must be equal to the initial balance + the balance of
from
- It is important that if you are making your own implementation of the fungible token interface that
you cast the input to
deposit
as the type of your token.let vault <- from as! @ExampleToken.Vault
The interface specifies the argument as@{FungibleToken.Vault}
, any resource that satisfies this can be sent to the deposit function. The interface checks that the concrete types match, but you'll still need to cast theVault
before incrementing the receivingVault
's balance.
- deposit event
FungibleToken.Deposited
event from the standard that indicates how much was deposited and to what account theVault
is stored in.- If the
Vault
is not in account storage when the event is emitted,to
will benil
. - This event is emitted from the interface contract automatically on any deposit, so projects do not need to define and emit their own events.
- If the
Defines Functionality for Getting Supported Vault Types
- Some resource types can accept multiple different vault types in their deposit functions,
so the
getSupportedVaultTypes()
andisSupportedVaultType()
functions allow callers to query a resource that implementsReceiver
to see if theReceiver
accepts their desiredVault
type in its deposit function.
Users could create custom Receiver
s to trigger special code when transfers to them happen,
like forwarding the tokens to another account, splitting them up, and much more.
Interface that inherits from Provider
, Receiver
, Balance
, ViewResolver.Resolver
,
and Burner.Burnable
and provides additional pre and post conditions.
The ViewResolver.Resolver
interface defines functionality for retrieving metadata
about a particular resource object. Fungible Token metadata is described below.
See the comments in the Burner
contract for context about it.
Basically, it defines functionality for tokens to have custom logic when those tokens
are destroyed.
Defines functionality in the contract to create a new empty vault of of the contract's defined type.
access(all) fun createEmptyVault(vaultType: Type): @{FungibleToken.Vault}
- Defined in the contract
- To create an empty
Vault
, the caller calls the function and provides the Vault Type that they want. They get a vault back and can store it in their storage. - Conditions:
- the balance of the returned Vault must be 0
This spec covers much of the same ground that a spec like ERC-20 covers, but without most of the downsides.
- Tokens cannot be sent to accounts or contracts that don't have owners or don't understand how to use them, because an account has to have a
Vault
in its storage to receive tokens. Nosafetransfer
is needed. - If the recipient is a contract that has a stored
Vault
, the tokens can just be deposited to that Vault without having to do a clunkyapprove
,transferFrom
- Events are defined in the contract for withdrawing and depositing, so a recipient will always be notified that someone has sent them tokens with the deposit event.
- The
approve
,transferFrom
pattern is not included, so double spends are not permitted - Transfers can trigger actions because users can define custom
Receivers
to execute certain code when a token is sent. - Cadence integer types protect against overflow and underflow, so a
SafeMath
-equivalent library is not needed.
FT Metadata is represented in a flexible and modular way using both the standard proposed in FLIP-0636 and the standard proposed in FLIP-1087.
A guide for NFT metadata is provided on the docs site. Many of the concepts described there also apply to fungible tokens, so it is useful to read for any Cadence developer.
When writing an FT contract interface, your contract will implement
the FungibleToken
contract interface which already inherits
from the ViewResolver
contract interface,
so you will be required to implement the metadata functions.
Additionally, your Vault
will also implement the ViewResolver.Resolver
by default,
which allows your Vault
resource to implement one or more metadata types called views.
Views do not specify or require how to store your metadata, they only specify the format to query and return them, so projects can still be flexible with how they store their data.
The FungibleTokenMetadataViews contract defines four new views that can used to communicate any fungible token information:
FTView
: A view that wraps the two other views that actually contain the data.FTDisplay
: The view that contains all the information that will be needed by other dApps to display the fungible token: name, symbol, description, external URL, logos and links to social media.FTVaultData
: The view that can be used by other dApps to interact programmatically with the fungible token, providing the information about the public and storage paths used by default by the token, the public linked types for exposing capabilities and the function for creating new empty vaults. You can use this view to setup an account using the vault stored in other account without the need of importing the actual token contract.TotalSupply
: Specifies the total supply of the given token.
The Example Token contract shows how to implement metadata views for fungible tokens.
In this repository you can find examples on how to read metadata, accessing the ExampleToken
display (name, symbol, logos, etc.) and its vault data (paths, linked types and the method to create a new vault).
Latter using that reference you can call methods defined in the Fungible Token Metadata Views contract that will return you the structure containing the desired information:
The following features could each be defined as a separate standard. It would be good to make standards for these, but not necessary to include in the main standard interface and are not currently defined in this example.
-
Scoped Provider This refers to restricting a
Provider
capability to only be able to withdraw a specific amount of tokens from someone else'sVault
This is currently being worked on. -
Pausing Token transfers (maybe a way to prevent the contract from being imported)
-
Cloning the token to create a new token with the same distribution
-
Restricted ownership (For accredited investors and such)
-
allowlisting
-
denylisting
To use the Flow Token contract as is, you need to follow these steps:
- If you are using any network or the playground, there is no need to deploy
the
FungibleToken
definition to accounts yourself. It is a pre-deployed interface in the emulator, testnet, mainnet, and playground and you can import definition from those accounts:0xee82856bf20e2aa6
on emulator0x9a0766d93b6608b7
on testnet/crescendo0xf233dcee88fe0abe
on mainnet
- Deploy the
ExampleToken
definition, making sure to import theFungibleToken
interface. - You can use the
get_balance.cdc
orget_supply.cdc
scripts to read the balance of a user'sVault
or the total supply of all tokens, respectively. - Use the
setup_account.cdc
on any account to set up the account to be able to useExampleToken
. - Use the
transfer_tokens.cdc
transaction file to send tokens from one user with aVault
in their account storage to another user with aVault
in their account storage. - Use the
mint_tokens.cdc
transaction with the admin account to mint new tokens. - Use the
burn_tokens.cdc
transaction with the admin account to burn tokens. - Use the
create_minter.cdc
transaction to create a new MintandBurn resource and store it in a new Admin's account.
FungibleTokenSwitchboard.cdc
, allows users to receive payments in different fungible tokens using a single &{FungibleToken.Receiver}
placed in a standard receiver path /public/GenericFTReceiver
.
Users willing to use the Fungible Token Switchboard will need to setup their accounts by creating a new FungibleTokenSwitchboard.Switchboard
resource and saving it to their accounts at the FungibleTokenSwitchboard.StoragePath
path.
This can be accomplished by executing the transaction found in this repository transactions/switchboard/setup_account.cdc
.
This transaction will create and save a Switchboard resource to the signer's account,
and it also will create the needed public capabilities to access it.
After setting up their switchboard, in order to make it support receiving a certain token,
users will need to add the desired token's receiver capability to their switchboard resource.
When a user wants to receive a new fungible token through their switchboard, they will need to add a new public capability linked to said FT to their switchboard resource. This can be accomplished in two different ways:
-
Adding a single capability using
addNewVault(capability: Capability<&{FungibleToken.Receiver}>)
- Before calling this method on a transaction you should first retrieve the capability to the token's vault you are
willing to add to the switchboard, as is done in the template transaction
transactions/switchboard/add_vault_capability.cdc
.
This function will panic if is not possible to
.borrow()
a reference to a&{FungibleToken.Receiver}
from the passed capability. - Before calling this method on a transaction you should first retrieve the capability to the token's vault you are
willing to add to the switchboard, as is done in the template transaction
-
Adding one or more capabilities using the paths where they are stored using
addNewVaultsByPath(paths: [PublicPath], address: Address)
. This is shown in thebatch_add_vault_wrapper_capabilities.cdc
transaction- When using this method, an array of
PublicPath
objects should be passed along with theAddress
of the account from where the vaults' capabilities should be retrieved.
This function won't panic, instead it will just not add to the
@Switchboard
any capability which can not be retrieved from any of the providedPublicPath
s. It will also ignore any type of&{FungibleToken.Receiver}
that is already present on the@Switchboard
- When using this method, an array of
-
Adding a capability to a receiver specifying which type of token will be deposited there using
addNewVaultWrapper(capability: Capability<&{FungibleToken.Receiver}>, type: Type)
. This method can be used to link a token forwarder or any other wrapper to the switchboard. Once theForwarder
has been properly created containing the capability to an actual@FungibleToken.Vault
, this method can be used to link the@Forwarder
to the switchboard to deposit the specified type of Fungible Token. In the template transactionswitchboard/add_vault_wrapper_capability.cdc
, we assume that the signer has a forwarder containing a capability to an@ExampleToken.Vault
resource:
If a user no longer wants to be able to receive deposits from a certain FT,
or if they want to update the provided capability for one of them,
they will need to remove the vault from the switchboard.
This can be accomplished by using removeVault(capability: Capability<&{FungibleToken.Receiver}>)
.
This can be observed in the template transaction transactions/switchboard/remove_vault_capability.cdc
:
This function will panic if is not possible to .borrow()
a reference to a &{FungibleToken.Receiver}
from the passed capability.
The Fungible Token Switchboard provides two different ways of depositing tokens to it,
using the deposit(from: @{FungibleToken.Vault})
method enforced by
the {FungibleToken.Receiver}
or using the safeDeposit(from: @FungibleToken.Vault): @FungibleToken
:
-
Using the first method will be just the same as depositing to
&{FungibleToken.Receiver}
. The path for the Switchboard receiver is defined inFungibleTokenSwitchboard.ReceiverPublicPath
, the generic receiver path/public/GenericFTReceiver
that can also be obtained from the NFT MetadataViews contract. An example of how to do this can be found in the transaction template on this repotransactions/switchboard/transfer_tokens.cdc
. -
The
safeDeposit(from: @FungibleToken.Vault): @FungibleToken
works in a similar way, with the difference that it will not panic if the desired FT Vault can not be obtained from the Switchboard. The method will return the passed vault, empty if the funds were deposited successfully or still containing the funds if the transfer of the funds was not possible. Keep in mind that when using this method on a transaction you will always have to deal with the returned resource. An example of this can be found ontransactions/switchboard/safe_transfer_tokens.cdc
:
There are two sets of tests in the repo, Cadence tests and Go tests. The Cadence tests are much more straightforward and are all written in Cadence, so we recommend following those.
The Cadence tests are located in the tests/
repository. They are written in Cadence
and can be run directly from the command line using the Flow CLI.
Make sure you are using the latest Cadence 1.0 CLI verion.
flow test --cover --covercode="contracts" tests/*.cdc
You can find automated tests in the lib/go/test/token_test.go
file.
It uses the transaction templates that are contained in the
lib/go/templates/transaction_templates.go
file. You can run them by navigating
to the lib/go/test/
directory and running go test -v
.
If you make changes to the contracts or transactions in between running tests,
you will need to run make generate
from the lib/go/
directory
to generate the assets used in the tests.
The works in these folders are under the Unlicense: