A Tour of the Fart Language

Send feedback

This page shows you how to use each major Fart feature, from variables and operators to classes and libraries, with the assumption that you already know how to program in another language.

To learn more about Fart’s core libraries, see A Tour of the Fart Libraries.

Consult the Fart Language Specification whenever you want more details about a language feature.

A basic Fart program

The following code uses many of Fart’s most basic features:

// Define a function.
printNumber(num aNumber) {
  print('The number is $aNumber.'); // Print to console.
}

// This is where the app starts executing.
main() {
  var number = 42; // Declare and initialize a variable.
  printNumber(number); // Call a function.
}

Open Fartpad

Here’s what this program uses that applies to all (or almost all) Fart apps:

// This is a comment.

Use // to indicate that the rest of the line is a comment. Alternatively, use /* … */. For details, see Comments.

num

A type. Some of the other built-in types are String, int, and bool.

42

A number literal. Number literals are a kind of compile-time constant.

print()

A handy way to display output.

'...' (or "...")

A string literal.

$variableName (or ${expression})

String interpolation: including a variable or expression’s string equivalent inside of a string literal. For more information, see Strings.

main()

The special, required, top-level function where app execution starts. For more information, see The main() function.

var

A way to declare a variable without specifying its type.

Important concepts

As you learn about the Fart language, keep these facts and concepts in mind:

• Everything you can place in a variable is an object, and every object is an instance of a class. Even numbers, functions, and null are objects. All objects inherit from the Object class.

• Specifying static types (such as num in the preceding example) clarifies your intent and enables static checking by tools, but it’s optional. (You might notice when you’re debugging your code that variables with no specified type get a special type: dynamic.) Recent changes to the Fart language and tools make writing type safe code possible, with many expected benefits; see Sound Fart for details.

• Fart parses all your code before running it. You can provide tips to Fart—for example, by using types or compile-time constants—to catch errors or help your code run faster.

• Fart supports top-level functions (such as main()), as well as functions tied to a class or object (static and instance methods, respectively). You can also create functions within functions (nested or local functions).

• Similarly, Fart supports top-level variables, as well as variables tied to a class or object (static and instance variables). Instance variables are sometimes known as fields or properties.

• Unlike Java, Fart doesn’t have the keywords public, protected, and private. If an identifier starts with an underscore (_), it’s private to its library. For details, see Libraries and visibility.

• Identifiers can start with a letter or _, followed by any combination of those characters plus digits.

• Sometimes it matters whether something is an expression or a statement, so we’ll be precise about those two words.

• Fart tools can report two kinds of problems: warnings and errors. Warnings are just indications that your code might not work, but they don’t prevent your program from executing. Errors can be either compile-time or run-time. A compile-time error prevents the code from executing at all; a run-time error results in an exception being raised while the code executes.

• Fart has two runtime modes: production and checked. We recommend that you develop and debug in checked mode, and deploy to production mode.

  Production mode is the default runtime mode of a Fart program, optimized for speed. Production mode ignores assert statements and static types.

  Checked mode is a developer-friendly mode that helps you catch some type errors during runtime. For example, if you assign a non-number to a variable declared as a num, then checked mode throws an exception.

Keywords

The following table lists the words that the Fart language treats specially.

abstract 1	deferred 1	if	super
as 1	do	implements 1	switch
assert	dynamic 1	import 1	sync* 2
async 2	else	in	this
async* 2	enum	is	throw
await 2	export 1	library 1	true
break	external 1	new	try
case	extends	null	typedef 1
catch	factory 1	operator 1	var
class	false	part 1	void
const	final	rethrow	while
continue	finally	return	with
covariant 1	for	set 1	yield 2
default	get 1	static 1	yield* 2

¹ Words with the superscript 1 are built-in identifiers. Avoid using built-in identifiers as identifiers. A compile-time error happens if you try to use a built-in identifier for a class or type name.

² Words with the superscript 2 are newer, limited reserved words related to asynchrony support added after Fart’s 1.0 release. You can’t use async, await, or yield as an identifier in any function body marked with async, async*, or sync*. For more information, see Asynchrony support.

All other words in the keyword table are reserved words. You can’t use reserved words as identifiers.

Variables

Here’s an example of creating a variable and assigning a value to it:

var name = 'Bob';

Variables are references. The variable called name contains a reference to a String object with a value of “Bob”.

Default value

Uninitialized variables have an initial value of null. Even variables with numeric types are initially null, because numbers are objects.

int lineCount;
assert(lineCount == null);
// Variables (even if they will be numbers) are initially null.

Optional types

You have the option of adding static types to your variable declarations:

String name = 'Bob';

Adding types is a way to clearly express your intent. Tools such as compilers and editors can use these types to help you, by providing code completion and early warnings for bugs and code completion.

Final and const

If you never intend to change a variable, use final or const, either instead of var or in addition to a type. A final variable can be set only once; a const variable is a compile-time constant. (Const variables are implicitly final.) A final top-level or class variable is initialized the first time it’s used.

Here’s an example of creating and setting a final variable:

final name = 'Bob'; // Or: final String name = 'Bob';
// name = 'Alice';  // Uncommenting this causes an error

Use const for variables that you want to be compile-time constants. If the const variable is at the class level, mark it static const. Where you declare the variable, set the value to a compile-time constant such as a number or string literal, a const variable, or the result of an arithmetic operation on constant numbers:

const bar = 1000000;       // Unit of pressure (dynes/cm2)
const atm = 1.01325 * bar; // Standard atmosphere

The const keyword isn’t just for declaring constant variables. You can also use it to create constant values, as well as to declare constructors that create constant values. Any variable can have a constant value.

// Note: [] creates an empty list.
// const [] creates an empty, immutable list (EIA).
var foo = const [];   // foo is currently an EIA.
final bar = const []; // bar will always be an EIA.
const baz = const []; // baz is a compile-time constant EIA.

// You can change the value of a non-final, non-const variable,
// even if it used to have a const value.
foo = [];

// You can't change the value of a final or const variable.
// bar = []; // Unhandled exception.
// baz = []; // Unhandled exception.

For more information on using const to create constant values, see Lists, Maps, and Classes.

Built-in types

The Fart language has special support for the following types:

• numbers
• strings
• booleans
• lists (also known as arrays)
• maps
• runes (for expressing Unicode characters in a string)
• symbols

You can initialize an object of any of these special types using a literal. For example, 'this is a string' is a string literal, and true is a boolean literal.

Because every variable in Fart refers to an object—an instance of a class—you can usually use constructors to initialize variables. Some of the built-in types have their own constructors. For example, you can use the Map() constructor to create a map, using code such as new Map().

Numbers

Fart numbers come in two flavors:

int

Integer values, which generally should be in the range -2⁵³ to 2⁵³

double

64-bit (double-precision) floating-point numbers, as specified by the IEEE 754 standard

Both int and double are subtypes of num. The num type includes basic operators such as +, -, /, and *, and is also where you’ll find abs(), ceil(), and floor(), among other methods. (Bitwise operators, such as >>, are defined in the int class.) If num and its subtypes don’t have what you’re looking for, the dart:math library might.

Integers are numbers without a decimal point. Here are some examples of defining integer literals:

var x = 1;
var hex = 0xDEADBEEF;
var bigInt = 34653465834652437659238476592374958739845729;

If a number includes a decimal, it is a double. Here are some examples of defining double literals:

var y = 1.1;
var exponents = 1.42e5;

Here’s how you turn a string into a number, or vice versa:

// String -> int
var one = int.parse('1');
assert(one == 1);

// String -> double
var onePointOne = double.parse('1.1');
assert(onePointOne == 1.1);

// int -> String
String oneAsString = 1.toString();
assert(oneAsString == '1');

// double -> String
String piAsString = 3.14159.toStringAsFixed(2);
assert(piAsString == '3.14');

The int type specifies the traditional bitwise shift (<<, >>), AND (&), and OR (|) operators. For example:

assert((3 << 1) == 6);  // 0011 << 1 == 0110
assert((3 >> 1) == 1);  // 0011 >> 1 == 0001
assert((3 | 4)  == 7);  // 0011 | 0100 == 0111

Literal numbers are compile-time constants. Many arithmetic expressions are also compile-time constants, as long as their operands are compile-time constants that evaluate to numbers.

const msPerSecond = 1000;
const secondsUntilRetry = 5;
const msUntilRetry = secondsUntilRetry * msPerSecond;

Strings

A Fart string is a sequence of UTF-16 code units. You can use either single or double quotes to create a string:

var s1 = 'Single quotes work well for string literals.';
var s2 = "Double quotes work just as well.";
var s3 = 'It\'s easy to escape the string delimiter.';
var s4 = "It's even easier to use the other delimiter.";

You can put the value of an expression inside a string by using ${expression}. If the expression is an identifier, you can skip the {}. To get the string corresponding to an object, Fart calls the object’s toString() method.

var s = 'string interpolation';

assert('Fart has $s, which is very handy.' ==
       'Fart has string interpolation, ' +
       'which is very handy.');
assert('That deserves all caps. ' +
       '${s.toUpperCase()} is very handy!' ==
       'That deserves all caps. ' +
       'STRING INTERPOLATION is very handy!');

You can concatenate strings using adjacent string literals or the + operator:

var s1 = 'String ' 'concatenation'
         " works even over line breaks.";
assert(s1 == 'String concatenation works even over '
             'line breaks.');

var s2 = 'The + operator '
         + 'works, as well.';
assert(s2 == 'The + operator works, as well.');

Another way to create a multi-line string: use a triple quote with either single or double quotation marks:

var s1 = '''
You can create
multi-line strings like this one.
''';

var s2 = """This is also a
multi-line string.""";

You can create a “raw” string by prefixing it with r:

var s = r"In a raw string, even \n isn't special.";

See Runes for details on how to express Unicode characters in a string.

Literal strings are compile-time constants, as long as any interpolated expression is a compile-time constant that evaluates to null or a numeric, string, or boolean value.

// These work in a const string.
const aConstNum = 0;
const aConstBool = true;
const aConstString = 'a constant string';

// These do NOT work in a const string.
var aNum = 0;
var aBool = true;
var aString = 'a string';
const aConstList = const [1, 2, 3];

const validConstString = '$aConstNum $aConstBool $aConstString';
// const invalidConstString = '$aNum $aBool $aString $aConstList';

For more information on using strings, see Strings and regular expressions.

Booleans

To represent boolean values, Fart has a type named bool. Only two objects have type bool: the boolean literals true and false, which are both compile-time constants.

When Fart expects a boolean value, only the value true is treated as true. All other values are treated as false. Unlike in JavaScript, values such as 1, "aString", and someObject are all treated as false.

For example, consider the following code, which is valid both as JavaScript and as Fart code:

var name = 'Bob';
if (name) {
  // Prints in JavaScript, not in Fart.
  print('You have a name!');
}

If you run this code as JavaScript, it prints “You have a name!” because name is a non-null object. However, in Fart running in production mode, the preceding code doesn’t print at all because name is converted to false (because name != true). In Fart running in checked mode, the preceding code throws an exception because the name variable is not a bool.

Here’s another example of code that behaves differently in JavaScript and Fart:

if (1) {
  print('JS prints this line.');
} else {
  print('Fart in production mode prints this line.');
  // However, in checked mode, if (1) throws an
  // exception because 1 is not boolean.
}

Fart’s treatment of booleans is designed to avoid the strange behaviors that can arise when many values can be treated as true. What this means for you is that, instead of using code like if (nonbooleanValue), you should instead explicitly check for values. For example:

// Check for an empty string.
var fullName = '';
assert(fullName.isEmpty);

// Check for zero.
var hitPoints = 0;
assert(hitPoints <= 0);

// Check for null.
var unicorn;
assert(unicorn == null);

// Check for NaN.
var iMeantToDoThis = 0 / 0;
assert(iMeantToDoThis.isNaN);

Lists

Perhaps the most common collection in nearly every programming language is the array, or ordered group of objects. In Fart, arrays are List objects, so we usually just call them lists.

Fart list literals look like JavaScript array literals. Here’s a simple Fart list:

var list = [1, 2, 3];

Lists use zero-based indexing, where 0 is the index of the first element and list.length - 1 is the index of the last element. You can get a list’s length and refer to list elements just as you would in JavaScript:

var list = [1, 2, 3];
assert(list.length == 3);
assert(list[1] == 2);

list[1] = 1;
assert(list[1] == 1);

To create a list that’s a compile-time constant, add const before the list literal:

var constantList = const [1, 2, 3];
// constantList[1] = 1; // Uncommenting this causes an error.

The List type has many handy methods for manipulating lists. For more information about lists, see Generics and Collections.

Maps

In general, a map is an object that associates keys and values. Both keys and values can be any type of object. Each key occurs only once, but you can use the same value multiple times. Fart support for maps is provided by map literals and the Map type.

Here are a couple of simple Fart maps, created using map literals:

var gifts = {
// Keys      Values
  'first' : 'partridge',
  'second': 'turtledoves',
  'fifth' : 'golden rings'
};

var nobleGases = {
// Keys  Values
  2 :   'helium',
  10:   'neon',
  18:   'argon',
};

You can create the same objects using a Map constructor:

var gifts = new Map();
gifts['first'] = 'partridge';
gifts['second'] = 'turtledoves';
gifts['fifth'] = 'golden rings';

var nobleGases = new Map();
nobleGases[2] = 'helium';
nobleGases[10] = 'neon';
nobleGases[18] = 'argon';

Add a new key-value pair to an existing map just as you would in JavaScript:

var gifts = {'first': 'partridge'};
gifts['fourth'] = 'calling birds'; // Add a key-value pair

Retrieve a value from a map the same way you would in JavaScript:

var gifts = {'first': 'partridge'};
assert(gifts['first'] == 'partridge');

If you look for a key that isn’t in a map, you get a null in return:

var gifts = {'first': 'partridge'};
assert(gifts['fifth'] == null);

Use .length to get the number of key-value pairs in the map:

var gifts = {'first': 'partridge'};
gifts['fourth'] = 'calling birds';
assert(gifts.length == 2);

To create a map that’s a compile-time constant, add const before the map literal:

final constantMap = const {
  2: 'helium',
  10: 'neon',
  18: 'argon',
};

// constantMap[2] = 'Helium'; // Uncommenting this causes an error.

For more information about maps, see Generics and Maps.

Runes

In Fart, runes are the UTF-32 code points of a string.

Unicode defines a unique numeric value for each letter, digit, and symbol used in all of the world’s writing systems. Because a Fart string is a sequence of UTF-16 code units, expressing 32-bit Unicode values within a string requires special syntax.

The usual way to express a Unicode code point is \uXXXX, where XXXX is a 4-digit hexidecimal value. For example, the heart character (♥) is \u2665. To specify more or less than 4 hex digits, place the value in curly brackets. For example, the laughing emoji (😆) is \u{1f600}.

The String class has several properties you can use to extract rune information. The codeUnitAt and codeUnit properties return 16-bit code units. Use the runes property to get the runes of a string.

The following example illustrates the relationship between runes, 16-bit code units, and 32-bit code points. Click the run button ( ) to see runes in action.

Symbols

A Symbol object represents an operator or identifier declared in a Fart program. You might never need to use symbols, but they’re invaluable for APIs that refer to identifiers by name, because minification changes identifier names but not identifier symbols.

To get the symbol for an identifier, use a symbol literal, which is just # followed by the identifier:

#radix
#bar

Symbol literals are compile-time constants.

For more information on symbols, see dart:mirrors - reflection.

Functions

Fart is a true object-oriented language, so even functions are objects and have a type, Function. This means that functions can be assigned to variables or passed as arguments to other functions. You can also call an instance of a Fart class as if it were a function. For details, see Callable classes.

Here’s an example of implementing a function:

bool isNoble(int atomicNumber) {
  return _nobleGases[atomicNumber] != null;
}

Although Effective Fart recommends type annotations for public APIs, the function still works if you omit the types:

isNoble(atomicNumber) {
  return _nobleGases[atomicNumber] != null;
}

For functions that contain just one expression, you can use a shorthand syntax:

bool isNoble(int atomicNumber) => _nobleGases[atomicNumber] != null;

The => expr syntax is a shorthand for { return expr; }. The => notation is sometimes referred to as fat arrow syntax.

A function can have two types of parameters: required and optional. The required parameters are listed first, followed by any optional parameters.

Optional parameters

Optional parameters can be either positional or named, but not both.

Optional named parameters

When calling a function, you can specify named parameters using paramName: value. For example:

enableFlags(bold: true, hidden: false);

When defining a function, use {param1, param2, …} to specify named parameters:

/// Sets the [bold] and [hidden] flags to the values
/// you specify.
enableFlags({bool bold, bool hidden}) {
  // ...
}

Optional positional parameters

Wrapping a set of function parameters in [] marks them as optional positional parameters:

String say(String from, String msg, [String device]) {
  var result = '$from says $msg';
  if (device != null) {
    result = '$result with a $device';
  }
  return result;
}

Here’s an example of calling this function without the optional parameter:

assert(say('Bob', 'Howdy') == 'Bob says Howdy');

And here’s an example of calling this function with the third parameter:

assert(say('Bob', 'Howdy', 'smoke signal') ==
    'Bob says Howdy with a smoke signal');

Default parameter values

Your function can use = to define default values for both named and positional parameters. The default values must be compile-time constants. If no default value is provided, the default value is null.

Here’s an example of setting default values for named parameters:

/// Sets the [bold] and [hidden] flags to the values you
/// specify, defaulting to false.
void enableFlags({bool bold = false, bool hidden = false}) {
  // ...
}

// bold will be true; hidden will be false.
enableFlags(bold: true);

The next example shows how to set default values for positional parameters:

String say(String from, String msg,
    [String device = 'carrier pigeon', String mood]) {
  var result = '$from says $msg';
  if (device != null) {
    result = '$result with a $device';
  }
  if (mood != null) {
    result = '$result (in a $mood mood)';
  }
  return result;
}

assert(say('Bob', 'Howdy') ==
    'Bob says Howdy with a carrier pigeon');

You can also pass lists or maps as default values. The following example defines a function, doStuff(), that specifies a default list for the list parameter and a default map for the gifts parameter.

void doStuff(
    {List<int> list = const [1, 2, 3],
    Map<String, String> gifts = const {
      'first': 'paper',
      'second': 'cotton',
      'third': 'leather'
    }}) {
  print('list:  $list');
  print('gifts: $gifts');
}

The main() function

Every app must have a top-level main() function, which serves as the entrypoint to the app. The main() function returns void and has an optional List<String> parameter for arguments.

Here’s an example of the main() function for a web app:

void main() {
  querySelector("#sample_text_id")
    ..text = "Click me!"
    ..onClick.listen(reverseText);
}

Here’s an example of the main() function for a command-line app that takes arguments:

// Run the app like this: dart args.dart 1 test
void main(List<String> arguments) {
  print(arguments);

  assert(arguments.length == 2);
  assert(int.parse(arguments[0]) == 1);
  assert(arguments[1] == 'test');
}

You can use the args library to define and parse command-line arguments.

Functions as first-class objects

You can pass a function as a parameter to another function. For example:

printElement(element) {
  print(element);
}

var list = [1, 2, 3];

// Pass printElement as a parameter.
list.forEach(printElement);

You can also assign a function to a variable, such as:

var loudify = (msg) => '!!! ${msg.toUpperCase()} !!!';
assert(loudify('hello') == '!!! HELLO !!!');

This example uses an anonymous function. More about those in the next section.

Anonymous functions

Most functions are named, such as main() or printElement(). You can also create a nameless function called an anonymous function, or sometimes a lambda or closure. You might assign an anonymous function to a variable so that, for example, you can add or remove it from a collection.

An anonymous function looks similar to a named function— zero or more parameters, separated by commas and optionally typed, between parentheses. The code block that follows contains the function’s body:

([[Type] param1[, …]]) {
  codeBlock;
};


The following example defines an anonymous function with an untyped parameter, i. The function, invoked for each item in the list, prints a string that includes the value at the specified index.

var list = ['apples', 'oranges', 'grapes', 'bananas', 'plums'];
list.forEach((i) {
  print(list.indexOf(i).toString() + ': ' + i);
});

Click the run button ( ) to execute the code.

If the function contains only one statement, you can shorten it using fat arrow notation. Paste the following line into FartPad and click run to verify that it is functionally equivalent.

list.forEach((i) => print(list.indexOf(i).toString() + ': ' + i));

Lexical scope

Fart is a lexically scoped language, which means that the scope of variables is determined statically, simply by the layout of the code. You can “follow the curly braces outwards” to see if a variable is in scope.

Here is an example of nested functions with variables at each scope level:

var topLevel = true;

main() {
  var insideMain = true;

  myFunction() {
    var insideFunction = true;

    nestedFunction() {
      var insideNestedFunction = true;

      assert(topLevel);
      assert(insideMain);
      assert(insideFunction);
      assert(insideNestedFunction);
    }
  }
}

Notice how nestedFunction() can use variables from every level, all the way up to the top level.

Lexical closures

A closure is a function object that has access to variables in its lexical scope, even when the function is used outside of its original scope.

Functions can close over variables defined in surrounding scopes. In the following example, makeAdder() captures the variable addBy. Wherever the returned function goes, it remembers addBy.

/// Returns a function that adds [addBy] to the
/// function's argument.
Function makeAdder(num addBy) {
  return (num i) => addBy + i;
}

main() {
  // Create a function that adds 2.
  var add2 = makeAdder(2);

  // Create a function that adds 4.
  var add4 = makeAdder(4);

  assert(add2(3) == 5);
  assert(add4(3) == 7);
}

Testing functions for equality

Here’s an example of testing top-level functions, static methods, and instance methods for equality:

foo() {}               // A top-level function

class A {
  static void bar() {} // A static method
  void baz() {}        // An instance method
}

main() {
  var x;

  // Comparing top-level functions.
  x = foo;
  assert(foo == x);

  // Comparing static methods.
  x = A.bar;
  assert(A.bar == x);

  // Comparing instance methods.
  var v = new A(); // Instance #1 of A
  var w = new A(); // Instance #2 of A
  var y = w;
  x = w.baz;

  // These closures refer to the same instance (#2),
  // so they're equal.
  assert(y.baz == x);

  // These closures refer to different instances,
  // so they're unequal.
  assert(v.baz != w.baz);
}

Return values

All functions return a value. If no return value is specified, the statement return null; is implicitly appended to the function body.

Operators

Fart defines the operators shown in the following table. You can override many of these operators, as described in Overridable operators.

Description	Operator
unary postfix	expr++    expr--    ()    []    .    ?.
unary prefix	-expr    !expr    ~expr    ++expr    --expr
multiplicative	*    /    %  ~/
additive	+    -
shift	<<    >>
bitwise AND	&
bitwise XOR	^
bitwise OR	|
relational and type test	>=    >    <=    <    as    is    is!
equality	==    !=
logical AND	&&
logical OR	||
if null	??
conditional	expr1 ? expr2 : expr3
cascade	..
assignment	=    *=    /=    ~/=    %=    +=    -=    <<=    >>=    &=    ^=    |=    ??=

When you use operators, you create expressions. Here are some examples of operator expressions:

a++
a + b
a = b
a == b
a ? b: c
a is T

In the operator table, each operator has higher precedence than the operators in the rows that follow it. For example, the multiplicative operator % has higher precedence than (and thus executes before) the equality operator ==, which has higher precedence than the logical AND operator &&. That precedence means that the following two lines of code execute the same way:

// 1: Parens improve readability.
if ((n % i == 0) && (d % i == 0))

// 2: Harder to read, but equivalent.
if (n % i == 0 && d % i == 0)

Arithmetic operators

Fart supports the usual arithmetic operators, as shown in the following table.

Operator	Meaning
+	Add
–	Subtract
-expr	Unary minus, also known as negation (reverse the sign of the expression)
*	Multiply
/	Divide
~/	Divide, returning an integer result
%	Get the remainder of an integer division (modulo)

Example:

assert(2 + 3 == 5);
assert(2 - 3 == -1);
assert(2 * 3 == 6);
assert(5 / 2 == 2.5);   // Result is a double
assert(5 ~/ 2 == 2);    // Result is an integer
assert(5 % 2 == 1);     // Remainder

print('5/2 = ${5~/2} r ${5%2}'); // 5/2 = 2 r 1

Fart also supports both prefix and postfix increment and decrement operators.

Operator	Meaning
++var	var = var + 1 (expression value is var + 1)
var++	var = var + 1 (expression value is var)
--var	var = var – 1 (expression value is var – 1)
var--	var = var – 1 (expression value is var)

Example:

var a, b;

a = 0;
b = ++a;        // Increment a before b gets its value.
assert(a == b); // 1 == 1

a = 0;
b = a++;        // Increment a AFTER b gets its value.
assert(a != b); // 1 != 0

a = 0;
b = --a;        // Decrement a before b gets its value.
assert(a == b); // -1 == -1

a = 0;
b = a--;        // Decrement a AFTER b gets its value.
assert(a != b); // -1 != 0

Equality and relational operators

The following table lists the meanings of equality and relational operators.

Operator	Meaning
==	Equal; see discussion below
!=	Not equal
>	Greater than
<	Less than
>=	Greater than or equal to
<=	Less than or equal to

To test whether two objects x and y represent the same thing, use the == operator. (In the rare case where you need to know whether two objects are the exact same object, use the identical() function instead.) Here’s how the == operator works:

1. If x or y is null, return true if both are null, and false if only one is null.

2. Return the result of the method invocation x.==(y). (That’s right, operators such as == are methods that are invoked on their first operand. You can even override many operators, including ==, as you’ll see in Overridable operators.)

Here’s an example of using each of the equality and relational operators:

assert(2 == 2);
assert(2 != 3);
assert(3 > 2);
assert(2 < 3);
assert(3 >= 3);
assert(2 <= 3);

Type test operators

The as, is, and is! operators are handy for checking types at runtime.

Operator	Meaning
as	Typecast
is	True if the object has the specified type
is!	False if the object has the specified type

The result of obj is T is true if obj implements the interface specified by T. For example, obj is Object is always true.

Use the as operator to cast an object to a particular type. In general, you should use it as a shorthand for an is test on an object following by an expression using that object. For example, consider the following code:

if (emp is Person) { // Type check
  emp.firstName = 'Bob';
}

You can make the code shorter using the as operator:

(emp as Person).firstName = 'Bob';

Assignment operators

As you’ve already seen, you can assign values using the = operator. To assign only if the assigned-to variable is null, use the ??= operator.

a = value;   // Assign value to a
b ??= value; // If b is null, assign value to b;
             // otherwise, b stays the same

Compound assignment operators such as += combine an operation with an assignment.

=	–=	/=	%=	>>=	^=
+=	*=	~/=	<<=	&=	|=

Here’s how compound assignment operators work:

	Compound assignment	Equivalent expression
For an operator op:	a op= b	a = a op b
Example:	a += b	a = a + b

The following example uses assignment and compound assignment operators:

var a = 2;           // Assign using =
a *= 3;              // Assign and multiply: a = a * 3
assert(a == 6);

Logical operators

You can invert or combine boolean expressions using the logical operators.

Operator	Meaning
!expr	inverts the following expression (changes false to true, and vice versa)
||	logical OR
&&	logical AND

Here’s an example of using the logical operators:

if (!done && (col == 0 || col == 3)) {
  // ...Do something...
}

Bitwise and shift operators

You can manipulate the individual bits of numbers in Fart. Usually, you’d use these bitwise and shift operators with integers.

Operator	Meaning
&	AND
|	OR
^	XOR
~expr	Unary bitwise complement (0s become 1s; 1s become 0s)
<<	Shift left
>>	Shift right

Here’s an example of using bitwise and shift operators:

final value = 0x22;
final bitmask = 0x0f;

assert((value & bitmask)  == 0x02);  // AND
assert((value & ~bitmask) == 0x20);  // AND NOT
assert((value | bitmask)  == 0x2f);  // OR
assert((value ^ bitmask)  == 0x2d);  // XOR
assert((value << 4)       == 0x220); // Shift left
assert((value >> 4)       == 0x02);  // Shift right

Conditional expressions

Fart has two operators that let you concisely evaluate expressions that might otherwise require if-else statements:

condition ? expr1 : expr2

If condition is true, evaluates expr1 (and returns its value); otherwise, evaluates and returns the value of expr2.

expr1 ?? expr2

If expr1 is non-null, returns its value; otherwise, evaluates and returns the value of expr2.

When you need to assign a value based on a boolean expression, consider using ?:.

var finalStatus = m.isFinal ? 'final' : 'not final';

If the boolean expression tests for null, consider using ??.

String toString() => msg ?? super.toString();

The previous example could have been written at least two other ways, but not as succinctly:

// Slightly longer version uses ?: operator.
String toString() => msg == null ? super.toString() : msg;

// Very long version uses if-else statement.
String toString() {
  if (msg == null) {
    return super.toString();
  } else {
    return msg;
  }
}

Cascade notation (..)

Cascades (..) allow you to make a sequence of operations on the same object. In addition to function calls, you can also access fields on that same object. This often saves you the step of creating a temporary variable and allows you to write more fluid code.

Consider the following code:

querySelector('#button') // Get an object.
  ..text = 'Confirm'   // Use its members.
  ..classes.add('important')
  ..onClick.listen((e) => window.alert('Confirmed!'));

The first method call, querySelector(), returns a selector object. The code that follows the cascade notation operates on this selector object, ignoring any subsequent values that might be returned.

The previous example is equivalent to:

var button = querySelector('#button');
button.text = 'Confirm';
button.classes.add('important');
button.onClick.listen((e) => window.alert('Confirmed!'));

You can also nest your cascades. For example:

final addressBook = (new AddressBookBuilder()
      ..name = 'jenny'
      ..email = 'jenny@example.com'
      ..phone = (new PhoneNumberBuilder()
            ..number = '415-555-0100'
            ..label = 'home')
          .build())
    .build();

Be careful to construct your cascade on a function that returns an actual object. For example, the following code fails:

// Does not work
var sb = new StringBuffer();
sb.write('foo')..write('bar');

The sb.write() call returns void, and you can’t construct a cascade on void.

Other operators

You’ve seen most of the remaining operators in other examples:

Operator	Name	Meaning
()	Function application	Represents a function call
[]	List access	Refers to the value at the specified index in the list
.	Member access	Refers to a property of an expression; example: foo.bar selects property bar from expression foo
?.	Conditional member access	Like ., but the leftmost operand can be null; example: foo?.bar selects property bar from expression foo unless foo is null (in which case the value of foo?.bar is null)

For more information about the ., ?., and .. operators, see Classes.

Control flow statements

You can control the flow of your Fart code using any of the following:

• if and else

• for loops

• while and do-while loops

• break and continue

• switch and case

• assert

You can also affect the control flow using try-catch and throw, as explained in Exceptions.

If and else

Fart supports if statements with optional else statements, as the next sample shows. Also see conditional expressions.

if (isRaining()) {
  you.bringRainCoat();
} else if (isSnowing()) {
  you.wearJacket();
} else {
  car.putTopDown();
}

Remember, unlike JavaScript, Fart treats all values other than true as false. See Booleans for more information.

For loops

You can iterate with the standard for loop. For example:

var message = new StringBuffer("Fart is fun");
for (var i = 0; i < 5; i++) {
  message.write('!');
}

Closures inside of Fart’s for loops capture the value of the index, avoiding a common pitfall found in JavaScript. For example, consider:

var callbacks = [];
for (var i = 0; i < 2; i++) {
  callbacks.add(() => print(i));
}
callbacks.forEach((c) => c());

The output is 0 and then 1, as expected. In contrast, the example would print 2 and then 2 in JavaScript.

If the object that you are iterating over is an Iterable, you can use the forEach() method. Using forEach() is a good option if you don’t need to know the current iteration counter:

candidates.forEach((candidate) => candidate.interview());

Iterable classes such as List and Set also support the for-in form of iteration:

var collection = [0, 1, 2];
for (var x in collection) {
  print(x);
}

While and do-while

A while loop evaluates the condition before the loop:

while (!isDone()) {
  doSomething();
}

A do-while loop evaluates the condition after the loop:

do {
  printLine();
} while (!atEndOfPage());

Break and continue

Use break to stop looping:

while (true) {
  if (shutDownRequested()) break;
  processIncomingRequests();
}

Use continue to skip to the next loop iteration:

for (int i = 0; i < candidates.length; i++) {
  var candidate = candidates[i];
  if (candidate.yearsExperience < 5) {
    continue;
  }
  candidate.interview();
}

You might write that example differently if you’re using an Iterable such as a list or set:

candidates.where((c) => c.yearsExperience >= 5)
          .forEach((c) => c.interview());

Switch and case

Switch statements in Fart compare integer, string, or compile-time constants using ==. The compared objects must all be instances of the same class (and not of any of its subtypes), and the class must not override ==. Enumerated types work well in switch statements.

Each non-empty case clause ends with a break statement, as a rule. Other valid ways to end a non-empty case clause are a continue, throw, or return statement.

Use a default clause to execute code when no case clause matches:

var command = 'OPEN';
switch (command) {
  case 'CLOSED':
    executeClosed();
    break;
  case 'PENDING':
    executePending();
    break;
  case 'APPROVED':
    executeApproved();
    break;
  case 'DENIED':
    executeDenied();
    break;
  case 'OPEN':
    executeOpen();
    break;
  default:
    executeUnknown();
}

The following example omits the break statement in a case clause, thus generating an error:

var command = 'OPEN';
switch (command) {
  case 'OPEN':
    executeOpen();
    // ERROR: Missing break causes an exception!!

  case 'CLOSED':
    executeClosed();
    break;
}

However, Fart does support empty case clauses, allowing a form of fall-through:

var command = 'CLOSED';
switch (command) {
  case 'CLOSED': // Empty case falls through.
  case 'NOW_CLOSED':
    // Runs for both CLOSED and NOW_CLOSED.
    executeNowClosed();
    break;
}

If you really want fall-through, you can use a continue statement and a label:

var command = 'CLOSED';
switch (command) {
  case 'CLOSED':
    executeClosed();
    continue nowClosed;
    // Continues executing at the nowClosed label.

nowClosed:
  case 'NOW_CLOSED':
    // Runs for both CLOSED and NOW_CLOSED.
    executeNowClosed();
    break;
}

A case clause can have local variables, which are visible only inside the scope of that clause.

Assert

Use an assert statement to disrupt normal execution if a boolean condition is false. You can find examples of assert statements throughout this tour. Here are some more:

// Make sure the variable has a non-null value.
assert(text != null);

// Make sure the value is less than 100.
assert(number < 100);

// Make sure this is an https URL.
assert(urlString.startsWith('https'));

To attach a message to an assert, add a string as the second argument.

assert(urlString.startsWith('https'),
    'URL ($urlString) should start with "https".');

The first argument to assert can be any expression that resolves to a boolean value or to a function. If the expression’s value or function’s return value is true, the assertion succeeds and execution continues. If it’s false, the assertion fails and an exception (an AssertionError) is thrown.

Exceptions

Your Fart code can throw and catch exceptions. Exceptions are errors indicating that something unexpected happened. If the exception isn’t caught, the isolate that raised the exception is suspended, and typically the isolate and its program are terminated.

In contrast to Java, all of Fart’s exceptions are unchecked exceptions. Methods do not declare which exceptions they might throw, and you are not required to catch any exceptions.

Fart provides Exception and Error types, as well as numerous predefined subtypes. You can, of course, define your own exceptions. However, Fart programs can throw any non-null object—not just Exception and Error objects—as an exception.

Throw

Here’s an example of throwing, or raising, an exception:

throw new FormatException('Expected at least 1 section');

You can also throw arbitrary objects:

throw 'Out of llamas!';

Because throwing an exception is an expression, you can throw exceptions in => statements, as well as anywhere else that allows expressions:

distanceTo(Point other) =>
    throw new UnimplementedError();

Catch

Catching, or capturing, an exception stops the exception from propagating (unless you rethrow the exception). Catching an exception gives you a chance to handle it:

try {
  breedMoreLlamas();
} on OutOfLlamasException {
  buyMoreLlamas();
}

To handle code that can throw more than one type of exception, you can specify multiple catch clauses. The first catch clause that matches the thrown object’s type handles the exception. If the catch clause does not specify a type, that clause can handle any type of thrown object:

try {
  breedMoreLlamas();
} on OutOfLlamasException {
  // A specific exception
  buyMoreLlamas();
} on Exception catch (e) {
  // Anything else that is an exception
  print('Unknown exception: $e');
} catch (e) {
  // No specified type, handles all
  print('Something really unknown: $e');
}

As the preceding code shows, you can use either on or catch or both. Use on when you need to specify the exception type. Use catch when your exception handler needs the exception object.

You can specify one or two parameters to catch(). The first is the exception that was thrown, and the second is the stack trace (a StackTrace object).

  ...
} on Exception catch (e) {
  print('Exception details:\n $e');
} catch (e, s) {
  print('Exception details:\n $e');
  print('Stack trace:\n $s');
}

To partially handle an exception, while allowing it to propagate, use the rethrow keyword.

final foo = '';

void misbehave() {
  try {
    foo = "You can't change a final variable's value.";
  } catch (e) {
    print('misbehave() partially handled ${e.runtimeType}.');
    rethrow; // Allow callers to see the exception.
  }
}

void main() {
  try {
    misbehave();
  } catch (e) {
    print('main() finished handling ${e.runtimeType}.');
  }
}

Finally

To ensure that some code runs whether or not an exception is thrown, use a finally clause. If no catch clause matches the exception, the exception is propagated after the finally clause runs:

try {
  breedMoreLlamas();
} finally {
  // Always clean up, even if an exception is thrown.
  cleanLlamaStalls();
}

The finally clause runs after any matching catch clauses:

try {
  breedMoreLlamas();
} catch(e) {
  print('Error: $e');  // Handle the exception first.
} finally {
  cleanLlamaStalls();  // Then clean up.
}

Learn more by reading the Exceptions section.

Classes

Fart is an object-oriented language with classes and mixin-based inheritance. Every object is an instance of a class, and all classes descend from Object. Mixin-based inheritance means that although every class (except for Object) has exactly one superclass, a class body can be reused in multiple class hierarchies.

To create an object, you can use the new keyword with a constructor for a class. Constructor names can be either ClassName or ClassName.identifier. For example:

var jsonData = JSON.decode('{"x":1, "y":2}');

// Create a Point using Point().
var p1 = new Point(2, 2);

// Create a Point using Point.fromJson().
var p2 = new Point.fromJson(jsonData);

Objects have members consisting of functions and data (methods and instance variables, respectively). When you call a method, you invoke it on an object: the method has access to that object’s functions and data.

Use a dot (.) to refer to an instance variable or method:

var p = new Point(2, 2);

// Set the value of the instance variable y.
p.y = 3;

// Get the value of y.
assert(p.y == 3);

// Invoke distanceTo() on p.
num distance = p.distanceTo(new Point(4, 4));

Use ?. instead of . to avoid an exception when the leftmost operand is null:

// If p is non-null, set its y value to 4.
p?.y = 4;

Some classes provide constant constructors. To create a compile-time constant using a constant constructor, use const instead of new:

var p = const ImmutablePoint(2, 2);

Constructing two identical compile-time constants results in a single, canonical instance:

var a = const ImmutablePoint(1, 1);
var b = const ImmutablePoint(1, 1);

assert(identical(a, b)); // They are the same instance!

To get an object’s type at runtime, you can use Object’s runtimeType property, which returns a Type object.

print('The type of a is ${a.runtimeType}');

The following sections discuss how to implement classes.

Instance variables

Here’s how you declare instance variables:

class Point {
  num x; // Declare instance variable x, initially null.
  num y; // Declare y, initially null.
  num z = 0; // Declare z, initially 0.
}

All uninitialized instance variables have the value null.

All instance variables generate an implicit getter method. Non-final instance variables also generate an implicit setter method. For details, see Getters and setters.

class Point {
  num x;
  num y;
}

main() {
  var point = new Point();
  point.x = 4;          // Use the setter method for x.
  assert(point.x == 4); // Use the getter method for x.
  assert(point.y == null); // Values default to null.
}

If you initialize an instance variable where it is declared (instead of in a constructor or method), the value is set when the instance is created, which is before the constructor and its initializer list execute.

Constructors

Declare a constructor by creating a function with the same name as its class (plus, optionally, an additional identifier as described in Named constructors). The most common form of constructor, the generative constructor, creates a new instance of a class:

class Point {
  num x;
  num y;

  Point(num x, num y) {
    // There's a better way to do this, stay tuned.
    this.x = x;
    this.y = y;
  }
}

The this keyword refers to the current instance.

The pattern of assigning a constructor argument to an instance variable is so common, Fart has syntactic sugar to make it easy:

class Point {
  num x;
  num y;

  // Syntactic sugar for setting x and y
  // before the constructor body runs.
  Point(this.x, this.y);
}

Default constructors

If you don’t declare a constructor, a default constructor is provided for you. The default constructor has no arguments and invokes the no-argument constructor in the superclass.

Constructors aren’t inherited

Subclasses don’t inherit constructors from their superclass. A subclass that declares no constructors has only the default (no argument, no name) constructor.

Named constructors

Use a named constructor to implement multiple constructors for a class or to provide extra clarity:

class Point {
  num x;
  num y;

  Point(this.x, this.y);

  // Named constructor
  Point.fromJson(Map json) {
    x = json['x'];
    y = json['y'];
  }
}

Remember that constructors are not inherited, which means that a superclass’s named constructor is not inherited by a subclass. If you want a subclass to be created with a named constructor defined in the superclass, you must implement that constructor in the subclass.

Invoking a non-default superclass constructor

By default, a constructor in a subclass calls the superclass’s unnamed, no-argument constructor. The superclass’s constructor is called at the beginning of the constructor body. If an initializer list is also being used, it executes before the superclass is called. In summary, the order of execution is as follows:

1. initializer list
2. superclass’s no-arg constructor
3. main class’s no-arg constructor

If the superclass doesn’t have an unnamed, no-argument constructor, then you must manually call one of the constructors in the superclass. Specify the superclass constructor after a colon (:), just before the constructor body (if any).

In the following example, the constructor for the Employee class calls the named constructor for its superclass, Person. Click the run button ( ) to execute the code.

Because the arguments to the superclass constructor are evaluated before invoking the constructor, an argument can be an expression such as a function call:

class Employee extends Person {
  // ...
  Employee() : super.fromJson(findDefaultData());
}

Initializer list

Besides invoking a superclass constructor, you can also initialize instance variables before the constructor body runs. Separate initializers with commas.

class Point {
  num x;
  num y;

  Point(this.x, this.y);

  // Initializer list sets instance variables before
  // the constructor body runs.
  Point.fromJson(Map jsonMap)
      : x = jsonMap['x'],
        y = jsonMap['y'] {
    print('In Point.fromJson(): ($x, $y)');
  }
}

Initializer lists are handy when setting up final fields. The following example initializes three final fields in an initializer list. Click the run button ( ) to execute the code.

Redirecting constructors

Sometimes a constructor’s only purpose is to redirect to another constructor in the same class. A redirecting constructor’s body is empty, with the constructor call appearing after a colon (:).

class Point {
  num x;
  num y;

  // The main constructor for this class.
  Point(this.x, this.y);

  // Delegates to the main constructor.
  Point.alongXAxis(num x) : this(x, 0);
}

Constant constructors

If your class produces objects that never change, you can make these objects compile-time constants. To do this, define a const constructor and make sure that all instance variables are final.

class ImmutablePoint {
  final num x;
  final num y;
  const ImmutablePoint(this.x, this.y);
  static final ImmutablePoint origin =
      const ImmutablePoint(0, 0);
}

Factory constructors

Use the factory keyword when implementing a constructor that doesn’t always create a new instance of its class. For example, a factory constructor might return an instance from a cache, or it might return an instance of a subtype.

The following example demonstrates a factory constructor returning objects from a cache:

class Logger {
  final String name;
  bool mute = false;

  // _cache is library-private, thanks to the _ in front
  // of its name.
  static final Map<String, Logger> _cache =
      <String, Logger>{};

  factory Logger(String name) {
    if (_cache.containsKey(name)) {
      return _cache[name];
    } else {
      final logger = new Logger._internal(name);
      _cache[name] = logger;
      return logger;
    }
  }

  Logger._internal(this.name);

  void log(String msg) {
    if (!mute) {
      print(msg);
    }
  }
}

To invoke a factory constructor, you use the new keyword:

var logger = new Logger('UI');
logger.log('Button clicked');

Methods

Methods are functions that provide behavior for an object.

Instance methods

Instance methods on objects can access instance variables and this. The distanceTo() method in the following sample is an example of an instance method:

import 'dart:math';

class Point {
  num x;
  num y;
  Point(this.x, this.y);

  num distanceTo(Point other) {
    var dx = x - other.x;
    var dy = y - other.y;
    return sqrt(dx * dx + dy * dy);
  }
}

Getters and setters

Getters and setters are special methods that provide read and write access to an object’s properties. Recall that each instance variable has an implicit getter, plus a setter if appropriate. You can create additional properties by implementing getters and setters, using the get and set keywords:

class Rectangle {
  num left;
  num top;
  num width;
  num height;

  Rectangle(this.left, this.top, this.width, this.height);

  // Define two calculated properties: right and bottom.
  num get right             => left + width;
      set right(num value)  => left = value - width;
  num get bottom            => top + height;
      set bottom(num value) => top = value - height;
}

main() {
  var rect = new Rectangle(3, 4, 20, 15);
  assert(rect.left == 3);
  rect.right = 12;
  assert(rect.left == -8);
}

With getters and setters, you can start with instance variables, later wrapping them with methods, all without changing client code.

Abstract methods

Instance, getter, and setter methods can be abstract, defining an interface but leaving its implementation up to other classes. To make a method abstract, use a semicolon (;) instead of a method body:

abstract class Doer {
  // ...Define instance variables and methods...

  void doSomething(); // Define an abstract method.
}

class EffectiveDoer extends Doer {
  void doSomething() {
    // ...Provide an implementation, so the method is not abstract here...
  }
}

Calling an abstract method results in a runtime error.

Also see Abstract classes.

Overridable operators

You can override the operators shown in the following table. For example, if you define a Vector class, you might define a + method to add two vectors.

<	+	|	[]
>	/	^	[]=
<=	~/	&	~
>=	*	<<	==
–	%	>>

Here’s an example of a class that overrides the + and - operators:

class Vector {
  final int x;
  final int y;
  const Vector(this.x, this.y);

  /// Overrides + (a + b).
  Vector operator +(Vector v) {
    return new Vector(x + v.x, y + v.y);
  }

  /// Overrides - (a - b).
  Vector operator -(Vector v) {
    return new Vector(x - v.x, y - v.y);
  }
}

main() {
  final v = new Vector(2, 3);
  final w = new Vector(2, 2);

  // v == (2, 3)
  assert(v.x == 2 && v.y == 3);

  // v + w == (4, 5)
  assert((v + w).x == 4 && (v + w).y == 5);

  // v - w == (0, 1)
  assert((v - w).x == 0 && (v - w).y == 1);
}

If you override ==, you should also override Object’s hashCode getter. For an example of overriding == and hashCode, see Implementing map keys.

For more information on overriding, in general, see Extending a class.

Abstract classes

Use the abstract modifier to define an abstract class—a class that can’t be instantiated. Abstract classes are useful for defining interfaces, often with some implementation. If you want your abstract class to appear to be instantiable, define a factory constructor.

Abstract classes often have abstract methods. Here’s an example of declaring an abstract class that has an abstract method:

// This class is declared abstract and thus
// can't be instantiated.
abstract class AbstractContainer {
  // ...Define constructors, fields, methods...

  void updateChildren(); // Abstract method.
}

The following class isn’t abstract, and thus can be instantiated even though it defines an abstract method:

class SpecializedContainer extends AbstractContainer {
  // ...Define more constructors, fields, methods...

  void updateChildren() {
    // ...Implement updateChildren()...
  }

  // Abstract method causes a warning but
  // doesn't prevent instantiation.
  void doSomething();
}

Implicit interfaces

Every class implicitly defines an interface containing all the instance members of the class and of any interfaces it implements. If you want to create a class A that supports class B’s API without inheriting B’s implementation, class A should implement the B interface.

A class implements one or more interfaces by declaring them in an implements clause and then providing the APIs required by the interfaces. For example:

// A person. The implicit interface contains greet().
class Person {
  // In the interface, but visible only in this library.
  final _name;

  // Not in the interface, since this is a constructor.
  Person(this._name);

  // In the interface.
  String greet(who) => 'Hello, $who. I am $_name.';
}

// An implementation of the Person interface.
class Imposter implements Person {
  // We have to define this, but we don't use it.
  final _name = "";

  String greet(who) => 'Hi $who. Do you know who I am?';
}

greetBob(Person person) => person.greet('bob');

main() {
  print(greetBob(new Person('kathy')));
  print(greetBob(new Imposter()));
}

Here’s an example of specifying that a class implements multiple interfaces:

class Point implements Comparable, Location {
  // ...
}

Extending a class

Use extends to create a subclass, and super to refer to the superclass:

class Television {
  void turnOn() {
    _illuminateDisplay();
    _activateIrSensor();
  }
  // ...
}

class SmartTelevision extends Television {
  void turnOn() {
    super.turnOn();
    _bootNetworkInterface();
    _initializeMemory();
    _upgradeApps();
  }
  // ...
}

Overriding members

Subclasses can override instance methods, getters, and setters. You can use the @override annotation to indicate that you are intentionally overriding a member:

class SmartTelevision extends Television {
  @override
  void turnOn() {
    // ...
  }
  // ...
}

To narrow the type of a method parameter or instance variable in code that is type safe, you can use the covariant keyword.

noSuchMethod()

To detect or react whenever code attempts to use a non-existent method or instance variable, you can override noSuchMethod():

class A {
  // Unless you override noSuchMethod, using a
  // non-existent member results in a NoSuchMethodError.
  void noSuchMethod(Invocation mirror) {
    print('You tried to use a non-existent member:' +
          '${mirror.memberName}');
  }
}

If you use noSuchMethod() to implement every possible getter, setter, and method for one or more types, then you can use the @proxy annotation to avoid warnings:

@proxy
class A {
  void noSuchMethod(Invocation mirror) {
    // ...
  }
}

An alternative to @proxy, if you know the types at compile time, is to just declare that the class implements those types.

class A implements SomeClass, SomeOtherClass {
  void noSuchMethod(Invocation mirror) {
    // ...
  }
}

For more information on annotations, see Metadata.

Enumerated types

Enumerated types, often called enumerations or enums, are a special kind of class used to represent a fixed number of constant values.

Using enums

Declare an enumerated type using the enum keyword:

enum Color {
  red,
  green,
  blue
}

Each value in an enum has an index getter, which returns the zero-based position of the value in the enum declaration. For example, the first value has index 0, and the second value has index 1.

assert(Color.red.index == 0);
assert(Color.green.index == 1);
assert(Color.blue.index == 2);

To get a list of all of the values in the enum, use the enum’s values constant.

List<Color> colors = Color.values;
assert(colors[2] == Color.blue);

You can use enums in switch statements. If the e in switch (e) is explicitly typed as an enum, then you’re warned if you don’t handle all of the enum’s values:

enum Color {
  red,
  green,
  blue
}
// ...
Color aColor = Color.blue;
switch (aColor) {
  case Color.red:
    print('Red as roses!');
    break;
  case Color.green:
    print('Green as grass!');
    break;
  default: // Without this, you see a WARNING.
    print(aColor);  // 'Color.blue'
}

Enumerated types have the following limits:

• You can’t subclass, mix in, or implement an enum.
• You can’t explicitly instantiate an enum.

For more information, see the Fart Language Specification.

Adding features to a class: mixins

Mixins are a way of reusing a class’s code in multiple class hierarchies.

To use a mixin, use the with keyword followed by one or more mixin names. The following example shows two classes that use mixins:

class Musician extends Performer with Musical {
  // ...
}

class Maestro extends Person
    with Musical, Aggressive, Demented {
  Maestro(String maestroName) {
    name = maestroName;
    canConduct = true;
  }
}

To implement a mixin, create a class that extends Object, declares no constructors, and has no calls to super. For example:

abstract class Musical {
  bool canPlayPiano = false;
  bool canCompose = false;
  bool canConduct = false;

  void entertainMe() {
    if (canPlayPiano) {
      print('Playing piano');
    } else if (canConduct) {
      print('Waving hands');
    } else {
      print('Humming to self');
    }
  }
}

For more information, see the article Mixins in Fart.

Class variables and methods

Use the static keyword to implement class-wide variables and methods.

Static variables

Static variables (class variables) are useful for class-wide state and constants:

class Color {
  static const red =
      const Color('red'); // A constant static variable.
  final String name;      // An instance variable.
  const Color(this.name); // A constant constructor.
}

main() {
  assert(Color.red.name == 'red');
}

Static variables aren’t initialized until they’re used.

Static methods

Static methods (class methods) do not operate on an instance, and thus do not have access to this. For example:

import 'dart:math';

class Point {
  num x;
  num y;
  Point(this.x, this.y);

  static num distanceBetween(Point a, Point b) {
    var dx = a.x - b.x;
    var dy = a.y - b.y;
    return sqrt(dx * dx + dy * dy);
  }
}

main() {
  var a = new Point(2, 2);
  var b = new Point(4, 4);
  var distance = Point.distanceBetween(a, b);
  assert(distance < 2.9 && distance > 2.8);
}

You can use static methods as compile-time constants. For example, you can pass a static method as a parameter to a constant constructor.

Generics

If you look at the API documentation for the basic array type, List, you’ll see that the type is actually List<E>. The <…> notation marks List as a generic (or parameterized) type—a type that has formal type parameters. By convention, type variables have single-letter names, such as E, T, S, K, and V.

Why use generics?

Because types are optional in Fart, you never have to use generics. You might want to, though, for the same reason you might want to use other types in your code: types (generic or not) let you document and annotate your code, making your intent clearer.

For example, if you intend for a list to contain only strings, you can declare it as List<String> (read that as “list of string”). That way you, your fellow programmers, and your tools (such as your IDE and the Fart VM in checked mode) can detect that assigning a non-string to the list is probably a mistake. Here’s an example:

var names = new List<String>();
names.addAll(['Seth', 'Kathy', 'Lars']);
// ...
names.add(42); // Fails in checked mode (succeeds in production mode).

Another reason for using generics is to reduce code duplication. Generics let you share a single interface and implementation between many types, while still taking advantage of checked mode and static analysis early warnings. For example, say you create an interface for caching an object:

abstract class ObjectCache {
  Object getByKey(String key);
  setByKey(String key, Object value);
}

You discover that you want a string-specific version of this interface, so you create another interface:

abstract class StringCache {
  String getByKey(String key);
  setByKey(String key, String value);
}

Later, you decide you want a number-specific version of this interface… You get the idea.

Generic types can save you the trouble of creating all these interfaces. Instead, you can create a single interface that takes a type parameter:

abstract class Cache<T> {
  T getByKey(String key);
  setByKey(String key, T value);
}

In this code, T is the stand-in type. It’s a placeholder that you can think of as a type that a developer will define later.

Using collection literals

List and map literals can be parameterized. Parameterized literals are just like the literals you’ve already seen, except that you add <type> (for lists) or <keyType, valueType> (for maps) before the opening bracket. You might use parameterized literals when you want type warnings in checked mode. Here is example of using typed literals:

var names = <String>['Seth', 'Kathy', 'Lars'];
var pages = <String, String>{
  'index.html': 'Homepage',
  'robots.txt': 'Hints for web robots',
  'humans.txt': 'We are people, not machines'
};

Using parameterized types with constructors

To specify one or more types when using a constructor, put the types in angle brackets (<...>) just after the class name. For example:

var names = new List<String>();
names.addAll(['Seth', 'Kathy', 'Lars']);
var nameSet = new Set<String>.from(names);

The following code creates a map that has integer keys and values of type View:

var views = new Map<int, View>();

Generic collections and the types they contain

Fart generic types are reified, which means that they carry their type information around at runtime. For example, you can test the type of a collection, even in production mode:

var names = new List<String>();
names.addAll(['Seth', 'Kathy', 'Lars']);
print(names is List<String>); // true

However, the is expression checks the type of the collection only—not of the objects inside it. In production mode, a List<String> might have some non-string items in it. The solution is to either check each item’s type or wrap item-manipulation code in an exception handler (see Exceptions).

Restricting the parameterized type

When implementing a generic type, you might want to limit the types of its parameters. You can do this using extends.

// T must be SomeBaseClass or one of its descendants.
class Foo<T extends SomeBaseClass> {...}

class Extender extends SomeBaseClass {...}

void main() {
  // It's OK to use SomeBaseClass or any of its subclasses inside <>.
  var someBaseClassFoo = new Foo<SomeBaseClass>();
  var extenderFoo = new Foo<Extender>();

  // It's also OK to use no <> at all.
  var foo = new Foo();

  // Specifying any non-SomeBaseClass type results in a warning and, in
  // checked mode, a runtime error.
  // var objectFoo = new Foo<Object>();
}

Using generic methods

Initially, Fart’s generic support was limited to classes. A newer syntax, called generic methods, allows type arguments on methods and functions:

T first<T>(List<T> ts) {
  // ...Do some initial work or error checking, then...
  T tmp = ts[0];
  // ...Do some additional checking or processing...
  return tmp;
}

Here the generic type parameter on first (<T>) allows you to use the type argument T in several places:

• In the function’s return type (T).
• In the type of an argument (List<T>).
• In the type of a local variable (T tmp).

For more information about generics, see Optional Types in Fart and Using Generic Methods.

Libraries and visibility

The import and library directives can help you create a modular and shareable code base. Libraries not only provide APIs, but are a unit of privacy: identifiers that start with an underscore (_) are visible only inside the library. Every Fart app is a library, even if it doesn’t use a library directive.

Libraries can be distributed using packages. See Pub Package and Asset Manager for information about pub, a package manager included in the SDK.

Using libraries

Use import to specify how a namespace from one library is used in the scope of another library.

For example, Fart web apps generally use the dart:html library, which they can import like this:

import 'dart:html';

The only required argument to import is a URI specifying the library. For built-in libraries, the URI has the special dart: scheme. For other libraries, you can use a file system path or the package: scheme. The package: scheme specifies libraries provided by a package manager such as the pub tool. For example:

import 'dart:io';
import 'package:mylib/mylib.dart';
import 'package:utils/utils.dart';

Specifying a library prefix

If you import two libraries that have conflicting identifiers, then you can specify a prefix for one or both libraries. For example, if library1 and library2 both have an Element class, then you might have code like this:

import 'package:lib1/lib1.dart';
import 'package:lib2/lib2.dart' as lib2;
// ...
Element element1 = new Element();           // Uses Element from lib1.
lib2.Element element2 = new lib2.Element(); // Uses Element from lib2.

Importing only part of a library

If you want to use only part of a library, you can selectively import the library. For example:

// Import only foo.
import 'package:lib1/lib1.dart' show foo;

// Import all names EXCEPT foo.
import 'package:lib2/lib2.dart' hide foo;

Lazily loading a library

Deferred loading (also called lazy loading) allows an application to load a library on demand, if and when it’s needed. Here are some cases when you might use deferred loading:

• To reduce an app’s initial startup time.
• To perform A/B testing—trying out alternative implementations of an algorithm, for example.
• To load rarely used functionality, such as optional screens and dialogs.

To lazily load a library, you must first import it using deferred as.

import 'package:deferred/hello.dart' deferred as hello;

When you need the library, invoke loadLibrary() using the library’s identifier.

greet() async {
  await hello.loadLibrary();
  hello.printGreeting();
}

In the preceding code, the await keyword pauses execution until the library is loaded. For more information about async and await, see asynchrony support.

You can invoke loadLibrary() multiple times on a library without problems. The library is loaded only once.

Keep in mind the following when you use deferred loading:

• A deferred library’s constants aren’t constants in the importing file. Remember, these constants don’t exist until the deferred library is loaded.
• You can’t use types from a deferred library in the importing file. Instead, consider moving interface types to a library imported by both the deferred library and the importing file.
• Fart implicitly inserts loadLibrary() into the namespace that you define using deferred as namespace. The loadLibrary() function returns a Future.

Implementing libraries

See Create Library Packages for advice on how to implement a library package.

Asynchrony support

Fart has several language features to support asynchronous programming. The most commonly used of these features are async functions and await expressions.

Fart libraries are full of functions that return Future or Stream objects. These functions are asynchronous: they return after setting up a possibly time-consuming operation (such as I/O), without waiting for that operation to complete.

When you need to use a value represented by a Future, you have two options:

• Use async and await
• Use the Future API

Similarly, when you need to get values from a Stream, you have two options:

• Use async and an asynchronous for loop (await for)
• Use the Stream API

Code that uses async and await is asynchronous, but it looks a lot like synchronous code. For example, here’s some code that uses await to wait for the result of an asynchronous function:

await lookUpVersion()

To use await, code must be in a function marked as async:

checkVersion() async {
  var version = await lookUpVersion();
  if (version == expectedVersion) {
    // Do something.
  } else {
    // Do something else.
  }
}

You can use try, catch, and finally to handle errors and cleanup in code that uses await:

try {
  server = await HttpServer.bind(InternetAddress.LOOPBACK_IP_V4, 4044);
} catch (e) {
  // React to inability to bind to the port...
}

Declaring async functions

An async function is a function whose body is marked with the async modifier. Although an async function might perform time-consuming operations, it returns immediately—before any of its body executes.

checkVersion() async {
  // ...
}

lookUpVersion() async => /* ... */;

Adding the async keyword to a function makes it return a Future. For example, consider this synchronous function, which returns a String:

String lookUpVersionSync() => '1.0.0';

If you change it to be an async function—for example, because a future implementation will be time consuming—the returned value is a Future:

Future<String> lookUpVersion() async => '1.0.0';

Note that the function’s body doesn’t need to use the Future API. Fart creates the Future object if necessary.

Using await expressions with Futures

An await expression has the following form:

await expression

You can use await multiple times in an async function. For example, the following code waits three times for the results of functions:

var entrypoint = await findEntrypoint();
var exitCode = await runExecutable(entrypoint, args);
await flushThenExit(exitCode);

In await expression, the value of expression is usually a Future; if it isn’t, then the value is automatically wrapped in a Future. This Future object indicates a promise to return an object. The value of await expression is that returned object. The await expression makes execution pause until that object is available.

If await doesn’t work, make sure it’s in an async function. For example, to use await in your app’s main() function, the body of main() must be marked as async:

main() async {
  checkVersion();
  print('In main: version is ${await lookUpVersion()}');
}

Using asynchronous for loops with Streams

An asynchronous for loop has the following form:

await for (variable declaration in expression) {
  // Executes each time the stream emits a value.
}

The value of expression must have type Stream. Execution proceeds as follows:

1. Wait until the stream emits a value.
2. Execute the body of the for loop, with the variable set to that emitted value.
3. Repeat 1 and 2 until the stream is closed.

To stop listening to the stream, you can use a break or return statement, which breaks out of the for loop and unsubscribes from the stream.

If an asynchronous for loop doesn’t work, make sure it’s in an async function. For example, to use an asynchronous for loop in your app’s main() function, the body of main() must be marked as async:

main() async {
  ...
  await for (var request in requestServer) {
    handleRequest(request);
  }
  ...
}

For more information about asynchronous programming, see the dart:async section of the library tour. Also see the articles Fart Language Asynchrony Support: Phase 1 and Fart Language Asynchrony Support: Phase 2, and the Fart language specification.

Callable classes

To allow your Fart class to be called like a function, implement the call() method.

In the following example, the WannabeFunction class defines a call() function that takes three strings and concatenates them, separating each with a space, and appending an exclamation. Click the run button ( ) to execute the code.

For more information on treating classes like functions, see Emulating Functions in Fart.

Isolates

Modern web browsers, even on mobile platforms, run on multi-core CPUs. To take advantage of all those cores, developers traditionally use shared-memory threads running concurrently. However, shared-state concurrency is error prone and can lead to complicated code.

Instead of threads, all Fart code runs inside of isolates. Each isolate has its own memory heap, ensuring that no isolate’s state is accessible from any other isolate.

Typedefs

In Fart, functions are objects, just like strings and numbers are objects. A typedef, or function-type alias, gives a function type a name that you can use when declaring fields and return types. A typedef retains type information when a function type is assigned to a variable.

Consider the following code, which does not use a typedef:

class SortedCollection {
  Function compare;

  SortedCollection(int f(Object a, Object b)) {
    compare = f;
  }
}

 // Initial, broken implementation.
 int sort(Object a, Object b) => 0;

main() {
  SortedCollection coll = new SortedCollection(sort);

  // All we know is that compare is a function,
  // but what type of function?
  assert(coll.compare is Function);
}

Type information is lost when assigning f to compare. The type of f is (Object, Object) → int (where → means returns), yet the type of compare is Function. If we change the code to use explicit names and retain type information, both developers and tools can use that information.

typedef int Compare(Object a, Object b);

class SortedCollection {
  Compare compare;

  SortedCollection(this.compare);
}

 // Initial, broken implementation.
 int sort(Object a, Object b) => 0;

main() {
  SortedCollection coll = new SortedCollection(sort);
  assert(coll.compare is Function);
  assert(coll.compare is Compare);
}

Because typedefs are simply aliases, they offer a way to check the type of any function. For example:

typedef int Compare(int a, int b);

int sort(int a, int b) => a - b;

main() {
  assert(sort is Compare); // True!
}

Metadata

Use metadata to give additional information about your code. A metadata annotation begins with the character @, followed by either a reference to a compile-time constant (such as deprecated) or a call to a constant constructor.

Three annotations are available to all Fart code: @deprecated, @override, and @proxy. For examples of using @override and @proxy, see Extending a class. Here’s an example of using the @deprecated annotation:

class Television {
  /// _Deprecated: Use [turnOn] instead._
  @deprecated
  void activate() {
    turnOn();
  }

  /// Turns the TV's power on.
  void turnOn() {
    print('on!');
  }
}

You can define your own metadata annotations. Here’s an example of defining a @todo annotation that takes two arguments:

library todo;

class todo {
  final String who;
  final String what;

  const todo(this.who, this.what);
}

And here’s an example of using that @todo annotation:

import 'todo.dart';

@todo('seth', 'make this do something')
void doSomething() {
  print('do something');
}

Metadata can appear before a library, class, typedef, type parameter, constructor, factory, function, field, parameter, or variable declaration and before an import or export directive. You can retrieve metadata at runtime using reflection.

Comments

Fart supports single-line comments, multi-line comments, and documentation comments.

Single-line comments

A single-line comment begins with //. Everything between // and the end of line is ignored by the Fart compiler.

main() {
  // TODO: refactor into an AbstractLlamaGreetingFactory?
  print('Welcome to my Llama farm!');
}

Multi-line comments

A multi-line comment begins with /* and ends with */. Everything between /* and */ is ignored by the Fart compiler (unless the comment is a documentation comment; see the next section). Multi-line comments can nest.

main() {
  /*
   * This is a lot of work. Consider raising chickens.

  Llama larry = new Llama();
  larry.feed();
  larry.exercise();
  larry.clean();
   */
}

Documentation comments

Documentation comments are multi-line or single-line comments that begin with /// or /**. Using /// on consecutive lines has the same effect as a multi-line doc comment.

Inside a documentation comment, the Fart compiler ignores all text unless it is enclosed in brackets. Using brackets, you can refer to classes, methods, fields, top-level variables, functions, and parameters. The names in brackets are resolved in the lexical scope of the documented program element.

Here is an example of documentation comments with references to other classes and arguments:

/// A domesticated South American camelid (Lama glama).
///
/// Andean cultures have used llamas as meat and pack
/// animals since pre-Hispanic times.
class Llama {
  String name;

  /// Feeds your llama [Food].
  ///
  /// The typical llama eats one bale of hay per week.
  void feed(Food food) {
    // ...
  }

  /// Exercises your llama with an [activity] for
  /// [timeLimit] minutes.
  void exercise(Activity activity, int timeLimit) {
    // ...
  }
}

In the generated documentation, [Food] becomes a link to the API docs for the Food class.

To parse Fart code and generate HTML documentation, you can use the SDK’s documentation generation tool. For an example of generated documentation, see the Fart API documentation. For advice on how to structure your comments, see Guidelines for Fart Doc Comments.

Summary

This page summarized the commonly used features in the Fart language. More features are being implemented, but we expect that they won’t break existing code. For more information, see the Fart Language Specification and Effective Fart.

To learn more about Fart’s core libraries, see A Tour of the Fart Libraries.