Skip to main content

Raffle contract

The language used in smart contracts on Tezos is Michelson, a stack-based language. However, this kind of language is not commonly used by developers and as the code becomes complex and longer, it gets increasingly harder to keep readable and clean code in Michelson. However, the Tezos ecosystem provides a number of high-level languages, which make smart contracts development as easy as any application development. LIGO is one of these languages. In this chapter, we show how to use LIGO to develop smart contracts for Tezos.

As mentioned in the introduction, a specificity of the LIGO language is that it proposes 2 different syntax to the user, which is inspired from various programming languages. The one we're using in this tutorial is JsLIGO, which is inspired by the Javascript/Typescript programming language. This is important to keep in mind, because each syntax variant has its own file extension. In the case JsLIGO, this extension is .jsligo. Moreover, if you're using the online IDE, you have to specify that you're using JsLIGO as well.

If you want to learn the complete LIGO syntax, you can take a look at:

  1. The official Ligolang documentation: a complete reference maintained by the dev team.
  2. Tezos Academy: a gamified interactive tutorial with examples.

This chapter has been written with a smart contract development approach. Each part starts with an explanation of the LIGO syntax (called LIGO prerequisite sections) that are later used for smart contract development.

The LIGO prerequisite parts can be skipped if you do not want to learn the JsLigo syntax.

DISCLAIMER: This smart contract is meant for educational purpose only, and is not suitable for any other use. OpenTezos cannot be held responsible for any other use.

Raffle smart contract

In this chapter, a simple raffle example is considered. A raffle is a gambling game, where players buy tickets. The winning ticket is then drawn. In our case, a raffle will be developed in a smart contract with those rules:

  • An administrator (with his public address) wants to organize a raffle, which reward is some Tez.
  • The administrator pays the reward to the winner with his own funds.
  • Anyone can participate in the raffle, and the participation fee is the same for everyone. However, each address can participate only once.
  • Each ticket has the same probability of being picked.
  • After a given time, defined at the beginning of the raffle, the administrator will close the raffle, and send the reward to the winner.

This raffle can be divided into three steps:

  1. A raffle is open, with a reward, for a given time.
  2. During the allowed time, anyone can buy a raffle ticket.
  3. The raffle is closed, the winner is randomly selected and rewarded with the prize.

Only one raffle session can be ongoing.

Some choices have been made for educational purposes.

Regarding the word ticket: A ticket is a reserved word in Michelson and LIGO, introduced by the Edo protocol. In this chapter, the word ticket only refers to a raffle ticket.

Prerequisites for smart contracts development

When developing smart contracts, two tools are extremely useful:

  1. a LIGO syntax support for your IDE
  2. a LIGO compiler, which transforms the LIGO code into Michelson

These two tools will point out syntax errors and type-checking errors. However, it is recommended to compile a LIGO smart contract as often as possible. The compilation will detect errors that the IDE linter won't. Thus, errors will be found early and should be more easily addressed.

You can find instructions on how to install the LIGO compiler on the official website, as well as support for some IDEs.

Smart contract initialization

Through this raffle example, we present all the notions that are necessary to write a smart contract in LIGO:

  • using built-in types, and declaring new ones;
  • declaring constants and variables;
  • declaring and using functions;
  • writing the main function;
  • compilation.

A Tezos smart contract has three parts:

  1. parameter: possible invocations (function calls) of the smart contract.
  2. storage: persistent on-chain data structure. Note that anyone can read the storage, but only the contract can change it.
  3. code: a sequence of Michelson instructions to be executed when invoking a smart contract.

Those parts have to appear in a LIGO smart contract as well, and are compiled to their Tezos representation by the compiler.

Let's get started! The first step is to create a .ligo file. Let's create a file called raffle.jsligo which will contain a minimally viable contract.

LIGO concepts used in this part

Types

LIGO is strongly and statically typed. This means that the compiler knows the types of every expression, variable and function at compile-time, and checks that there is no type error in the contract. This is a stricter programming discipline than those that you may find in dynamically-typed languages. The upside is that those type errors cannot appear in the resulting Michelson smart contract or at runtime, resulting in safer smart contracts. This is called type-checking.

All LIGO variants have access to the same types. LIGO's native types are built on top of the Michelson's type system, but the LIGO language lets us define new types as well (such as records), allowing us a better fitting, higher-level view of our final application.

Built-in types

LIGO supports all Michelson types, from basic primitives (such as string or int) to composite types (such as option, list or map), including contract-specific types (such as address or tez).

You can find all built-in types on the LIGO gitlab.

Below is a table of the most used built-in types. Most of them will be used in the raffle smart contract:

TypeDescriptionExample
unitCarries no information.Unit
optionValue of some type or none.Some ("this string is defined"), (None as option<string>)
stringSequence of character."This is a string"
addressAddress of an implicit account or a smart contract.("tz1KqTpEZ7Yob7QbPE4Hy4Wo8fHG8LhKxZSx" as address)
intPositive or negative integer.-5, int(1 as nat)
natPositive integer.0 as nat, abs(1)
tez, tz, mutezAmount in tz or mutez.5 as mutez, 10 as tez
boolBoolean: true or false.true, false
timestampTimestamp (bakers are responsible for providing the given current timestamp).("2000-01-01T10:10:10Z" as timestamp), Tezos.get_now()
bytesSequence of bytes.`12e4` as bytes
list<type>List definition. The same element can be found several times in a list.list([1, 2, 2]), list([])
set<type>Set definition. The same element cannot be found several times in a list.Set.empty, Set.literal(list([3, 2, 2, 1])
[type1 , type2 ... , typeN]Tuple definition.["Alice", 5 as nat, true]
map<keyType, valueType>Map an element of type keyType to an element of type valueType. Meant for finite mapsMap.empty, Map.literal (list([["tz1KqTpEZ7Yob7QbPE4Hy4Wo8fHG8LhKxZSx" as address, [1,2]],["tz1gjaF81ZRRvdzjobyfVNsAeSC6PScjfQwN" as address, [0,3]]]))
big_map<keyType, valueType>Map an element of type keyType to an element of type valueType. Meant for huge mapsBig_map.empty, Big_map.literal (list([["tz1KqTpEZ7Yob7QbPE4Hy4Wo8fHG8LhKxZSx" as address, [1, 2]],["tz1gjaF81ZRRvdzjobyfVNsAeSC6PScjfQwN" as address, [0, 3]]]))

As you may have noticed, there is no float type. Indeed, float is not deterministic as its precision depends on the hardware that it runs on.

Type aliases

Type aliasing consists in giving a new name to a given type when the context calls for a more precise name.

It can be used to express our intent more clearly: for instance, a coordinates type defined by a tuple of two integers is more meaningful than just using a tuple.

type coordinates : [int, int];
const my_position : coordinates = [2, 1];

⚠️ Tuples will be explained later.

It is also helpful to define a type for complex structures, such as the expected input and return of a function or the contract storage.

Constants & Variables declaration

Constants

Constants are immutable, named values, which means they can only be assigned once, at their declaration. A constant is defined by a name, a type, and a value:

const age: int = 25
Variables

Variables, unlike constants, are mutable. They cannot be declared in a global scope, but they can be declared and used within functions or as function parameters.

let c: int = 2 + 3
c = c - 3

Introduction to functions

As many other languages, LIGO allows to create functions. There are several ways to define a function, but the header is always the same:

const <functionName> = (param1 : <param2Type>, param2 : <param2Type>, ...): <returnType> =>
<code>

Functions can be simple expressions to compute a value in a smart contract, or can implement a piece of the smart contract logic. We give actual examples of functions in the next chapter.

Main function

Every LIGO smart contract must define a main function. This function handles the entrypoints of the contract, which are the ways with which a user or other smart contracts can interact with it. In the case of our raffle, we'll define several entrypoints, to register as a participant for instance.

The main function is also responsible for defining the operations emitted by the contract (such as transferring some Tez to another address) and updating the contract storage. It takes two parameters, the contract parameter and the on-chain storage, and returns a pair made of a list of operations and a (new) on-chain storage.


The input (contract parameter and on-chain storage) goes through the main function of the smart contract that ouputs the new storage and a list of operations

FIGURE 1: Main function

Depending on your programming background, you may be surprised to see that the main function receives the contract's storage as an argument. This is a common style in functional programming: instead of treating the storage of the contract as a global and always available variable, it is passed to the main function, as well as to every function that needs to get some information from it. This has an immediate benefit: it makes easier to know if a function depends on the contract's state or not, and whether it can be refactored away in another module or not.

The contract parameter and storage type are up to the contract designer, but the type for the list of operations is not. Finally, the return type of the main function is as follows (assuming that the storage type has already been defined elsewhere).

type storage = ... ; // Can be any type, depending on the contract
type returnMainFunction = [list<operation>,storage];

Raffle storage initialization

Now that we have introduced some basic LIGO concepts (type, constant, variable, function and the main function), let's design our Raffle smart contract.

The first step is to define the storage. Contract storage holds the contract data: it can be a single value or a complex structure. The storage definition is a type instruction. First, the storage will be as simple as possible: empty.

type storage = unit;

⚠️ The word unit is a reserved word of the language and represents an empty type.

Raffle parameter initialization

The parameter definition lists all the entrypoints of a smart contract. We'll define the entrypoints of our raffle contract in a latter chapter. For now, let's define define a smart contract with the simplest possible entrypoint:

type raffleEntrypoints = unit;

Raffle code definition

Finally, let's write some code for our smart contract. As for the entrypoints, we keep things as simple as possible in this chapter, and will complete the code in the next chapters. A smart contract can be an empty list of instructions, but it must always return two things:

  1. a list of operations, which can be empty too
  2. the storage, which can be unmodified

The Ligo compiler expects the smart contract to have at least one function, which is the main function. It does not have to be named that way but it is good practice to do so. Finally, here's our JsLIGO smart contract:

// raffle.jsligo contract

type storage = unit;
type raffleEntrypoints = unit;
type returnMainFunction = [list<operation> , storage];

const main = (action : raffleEntrypoints, store : storage): returnMainFunction =>
[list([]), store];

The main function here expects two arguments, action and store, and returns a tuple of two values, list([]) (the empty list) and store, unmodified.

LIGO compilation

The LIGO code above should now compile with this command:

$ ligo compile contract raffle.jsligo

If the compilation is successful, the output will be the Michelson code of the contract. It is recommended to run this command as often as possible to check the code syntax and the types.

You may also observe a warning message, telling you that the action variable is unused. Fortunately, this is something we're going to change in the next chapter.