目录

修复常见的类型问题

If you're having problems with type checks, this page can help. To learn more, read about Dart's type system, and see these other resources.

Troubleshooting

#

Dart enforces a sound type system. This means you can't write code where a variable's value differs from its static type. A variable with an int type can't store a number with a decimal place. Dart checks variable values against their types at compile-time and runtime.

You can't get into a situation where the value stored in a variable is different from the variable's static type. Like most modern statically typed languages, Dart accomplishes this with a combination of static (compile-time) and dynamic (runtime) checking.

For example, the following type error is detected at compile-time:

✗ static analysis: failuredart
List<int> numbers = [1, 2, 3];
List<String> string = numbers;

Since neither List<int> nor List<String> is a subtype of the other, Dart rules this out statically.

You can see other examples of static analysis errors, as well as other error types, in the following sections.

No type errors

#

If you're not seeing expected errors or warnings, make sure that you're using the latest version of Dart and you have properly configured your IDE or editor.

You can also run analysis on your program using the command line with the dart analyze command.

To verify that analysis is working as expected, try adding the following code to a Dart file.

✗ static analysis: failuredart
bool b = [0][0];

If properly configured, the analyzer produces the following error:

error - A value of type 'int' can't be assigned to a variable of type 'bool'. Try changing the type of the variable, or casting the right-hand type to 'bool'. - invalid_assignment

Static errors and warnings

#

This section shows how to fix some of the errors and warnings you might see from the analyzer or an IDE.

Static analysis can't catch all errors. For help fixing errors that appear only at runtime, see Runtime errors.

Undefined member

#
error - The <member> '...' isn't defined for the type '...' - undefined_<member>

These errors can appear under the following conditions:

  • A variable is statically known to be some supertype, but the code assumes a subtype.
  • A generic class has a bounded type parameter, but an instance creation expression of the class omits the type argument.

Example 1: A variable is statically known to be some supertype, but the code assumes a subtype

#

In the following code, the analyzer complains that context2D is undefined:

✗ static analysis: failuredart
var canvas = querySelector('canvas')!;
canvas.context2D.lineTo(x, y);
error - The getter 'context2D' isn't defined for the type 'Element'. Try importing the library that defines 'context2D', correcting the name to the name of an existing getter, or defining a getter or field named 'context2D'. - undefined_getter

Fix: Replace the definition of the member with an explicit type declaration or a downcast

#

The return type of querySelector() is Element? (which the ! converts to Element), but the code assumes that it's the subtype CanvasElement (which defines context2D). The canvas field is declared as var, which allows Dart to infer canvas to be an Element.

You can fix this error with an explicit downcast:

✔ static analysis: successdart
var canvas = querySelector('canvas') as CanvasElement;
canvas.context2D.lineTo(x, y);

Otherwise, use dynamic in situations where you cannot use a single type:

✔ static analysis: successdart
dynamic canvasOrImg = querySelector('canvas, img');
var width = canvasOrImg.width;

Example 2: Omitted type parameters default to their type bounds

#

Consider the following generic class with a bounded type parameter that extends Iterable:

dart
class C<T extends Iterable> {
  final T collection;
  C(this.collection);
}

The following code creates a new instance of this class (omitting the type argument) and accesses its collection member:

✗ static analysis: failuredart
var c = C(Iterable.empty()).collection;
c.add(2);
error - The method 'add' isn't defined for the type 'Iterable'. Try correcting the name to the name of an existing method, or defining a method named 'add'. - undefined_method

While the List type has an add() method, Iterable does not.

Fix: Specify type arguments or fix downstream errors

#

When a generic class is instantiated without explicit type arguments, each type parameter defaults to its type bound (Iterable in this example) if one is explicitly given, or dynamic otherwise.

You need to approach fixing such errors on a case-by-case basis. It helps to have a good understanding of the original design intent.

Explicitly passing type arguments is an effective way to help identify type errors. For example, if you change the code to specify List as a type argument, the analyzer can detect the type mismatch in the constructor argument. Fix the error by providing a constructor argument of the appropriate type, such as a list literal:

✔ static analysis: successdart
var c = C<List>([]).collection;
c.add(2);

Invalid method override

#
error - '...'  isn't a valid override of '...' - invalid_override

These errors typically occur when a subclass tightens up a method's parameter types by specifying a subclass of the original class.

Example

#

In the following example, the parameters to the add() method are of type int, a subtype of num, which is the parameter type used in the parent class.

✗ static analysis: failuredart
abstract class NumberAdder {
  num add(num a, num b);
}

class MyAdder extends NumberAdder {
  @override
  num add(int a, int b) => a + b;
}
error - 'MyAdder.add' ('num Function(int, int)') isn't a valid override of 'NumberAdder.add' ('num Function(num, num)'). - invalid_override

Consider the following scenario where floating point values are passed to an MyAdder:

✗ runtime: failuredart
NumberAdder adder = MyAdder();
adder.add(1.2, 3.4);

If the override were allowed, the code would raise an error at runtime.

Fix: Widen the method's parameter types

#

The subclass's method should accept every object that the superclass's method takes.

Fix the example by widening the types in the subclass:

✔ static analysis: successdart
abstract class NumberAdder {
  num add(num a, num b);
}

class MyAdder extends NumberAdder {
  @override
  num add(num a, num b) => a + b;
}

For more information, see Use proper input parameter types when overriding methods.


Missing type arguments

#
error - '...'  isn't a valid override of '...' - invalid_override

Example

#

In the following example, Subclass extends Superclass<T> but doesn't specify a type argument. The analyzer infers Subclass<dynamic>, which results in an invalid override error on method(int).

✗ static analysis: failuredart
class Superclass<T> {
  void method(T param) { ... }
}

class Subclass extends Superclass {
  @override
  void method(int param) { ... }
}
error - 'Subclass.method' ('void Function(int)') isn't a valid override of 'Superclass.method' ('void Function(dynamic)'). - invalid_override

Fix: Specify type arguments for the generic subclass

#

When a generic subclass neglects to specify a type argument, the analyzer infers the dynamic type. This is likely to cause errors.

You can fix the example by specifying the type on the subclass:

✔ static analysis: successdart
class Superclass<T> {
  void method(T param) { ... }
}

class Subclass extends Superclass<int> {
  @override
  void method(int param) { ... }
}

Consider using the analyzer in strict raw types mode, which ensures that your code specifies generic type arguments. Here's an example of enabling strict raw types in your project's analysis_options.yaml file:

yaml
analyzer:
  language:
    strict-raw-types: true

To learn more about customizing the analyzer's behavior, see Customizing static analysis.


Unexpected collection element type

#
error - A value of type '...' can't be assigned to a variable of type '...' - invalid_assignment

This sometimes happens when you create a simple dynamic collection and the analyzer infers the type in a way you didn't expect. When you later add values of a different type, the analyzer reports an issue.

Example

#

The following code initializes a map with several (String, int) pairs. The analyzer infers that map to be of type <String, int> but the code seems to assume either <String, dynamic> or <String, num>. When the code adds a (String, double) pair, the analyzer complains:

✗ static analysis: failuredart
// Inferred as Map<String, int>
var map = {'a': 1, 'b': 2, 'c': 3};
map['d'] = 1.5;
error - A value of type 'double' can't be assigned to a variable of type 'int'. Try changing the type of the variable, or casting the right-hand type to 'int'. - invalid_assignment

Fix: Specify the type explicitly

#

The example can be fixed by explicitly defining the map's type to be <String, num>.

✔ static analysis: successdart
var map = <String, num>{'a': 1, 'b': 2, 'c': 3};
map['d'] = 1.5;

Alternatively, if you want this map to accept any value, specify the type as <String, dynamic>.


Constructor initialization list super() call

#
error - The superconstructor call must be last in an initializer list: '...'. - super_invocation_not_last

This error occurs when the super() call is not last in a constructor's initialization list.

Example

#
✗ static analysis: failuredart
HoneyBadger(Eats food, String name)
    : super(food),
      _name = name { ... }
error - The superconstructor call must be last in an initializer list: 'Animal'. - super_invocation_not_last

Fix: Put the super() call last

#

The compiler can generate simpler code if it relies on the super() call appearing last.

Fix this error by moving the super() call:

✔ static analysis: successdart
HoneyBadger(Eats food, String name)
    : _name = name,
      super(food) { ... }

The argument type ... can't be assigned to the parameter type ...

#
error - The argument type '...' can't be assigned to the parameter type '...'. - argument_type_not_assignable

In Dart 1.x dynamic was both a top type (supertype of all types) and a bottom type (subtype of all types) depending on the context. This meant it was valid to assign, for example, a function with a parameter of type String to a place that expected a function type with a parameter of dynamic.

However, in Dart 2 using a parameter type other than dynamic (or another top type, such as Object?) results in a compile-time error.

Example

#
✗ static analysis: failuredart
void filterValues(bool Function(dynamic) filter) {}
filterValues((String x) => x.contains('Hello'));
error - The argument type 'bool Function(String)' can't be assigned to the parameter type 'bool Function(dynamic)'. - argument_type_not_assignable

Fix: Add type parameters or cast from dynamic explicitly

#

When possible, avoid this error by adding type parameters:

✔ static analysis: successdart
void filterValues<T>(bool Function(T) filter) {}
filterValues<String>((x) => x.contains('Hello'));

Otherwise use casting:

✔ static analysis: successdart
void filterValues(bool Function(dynamic) filter) {}
filterValues((x) => (x as String).contains('Hello'));

Incorrect type inference

#

On rare occasions, Dart's type inference might infer the wrong type for function literal arguments in a generic constructor invocation. This primarily affects Iterable.fold.

Example

#

In the following code, type inference will infer that a has a type of Null:

✗ static analysis: failuredart
var ints = [1, 2, 3];
var maximumOrNull = ints.fold(null, (a, b) => a == null || a < b ? b : a);

Fix: Supply appropriate type as explicit type argument

#
✔ static analysis: successdart
var ints = [1, 2, 3];
var maximumOrNull =
    ints.fold<int?>(null, (a, b) => a == null || a < b ? b : a);

Runtime errors

#

The errors discussed in this section are reported at runtime.

Invalid casts

#

To ensure type safety, Dart needs to insert runtime checks in some cases. Consider the following assumeStrings method:

✔ static analysis: successdart
void assumeStrings(dynamic objects) {
  List<String> strings = objects; // Runtime downcast check
  String string = strings[0]; // Expect a String value
}

The assignment to strings is downcasting the dynamic to List<String> implicitly (as if you wrote as List<String>), so if the value you pass in objects at runtime is a List<String>, then the cast succeeds.

Otherwise, the cast will fail at runtime:

✗ runtime: failuredart
assumeStrings(<int>[1, 2, 3]);
Exception: type 'List<int>' is not a subtype of type 'List<String>'

Fix: Tighten or correct types

#

Sometimes, lack of a type, especially with empty collections, means that a <dynamic> collection is created, instead of the typed one you intended. Adding an explicit type argument can help:

runtime-successdart
var list = <String>[];
list.add('a string');
list.add('another');
assumeStrings(list);

You can also more precisely type the local variable, and let inference help:

runtime-successdart
List<String> list = [];
list.add('a string');
list.add('another');
assumeStrings(list);

In cases where you are working with a collection that you don't create, such as from JSON or an external data source, you can use the cast() method provided by Iterable implementations, such as List.

Here's an example of the preferred solution: tightening the object's type.

runtime-successdart
Map<String, dynamic> json = fetchFromExternalSource();
var names = json['names'] as List;
assumeStrings(names.cast<String>());

Appendix

#

The covariant keyword

#

Some (rarely used) coding patterns rely on tightening a type by overriding a parameter's type with a subtype, which is invalid. In this case, you can use the covariant keyword to tell the analyzer that you are doing this intentionally. This removes the static error and instead checks for an invalid argument type at runtime.

The following shows how you might use covariant:

✔ static analysis: successdart
class Animal {
  void chase(Animal x) { ... }
}

class Mouse extends Animal { ... }

class Cat extends Animal {
  @override
  void chase(covariant Mouse x) { ... }
}

Although this example shows using covariant in the subtype, the covariant keyword can be placed in either the superclass or the subclass method. Usually the superclass method is the best place to put it. The covariant keyword applies to a single parameter and is also supported on setters and fields.