I saw many people's code use async keyword in a function without an await keyword in the function body. Even some official flutter example code do this. I have no idea why. What is the point? Is this a mistake or having a purpose?
Normally, I just remove the async keyword from those code and everything will run without any problems. Can some dart expert clarify that if there is a purpose for a function which has the async keyword but NO await keyword? Or is this just their mistake?
async is sometimes used to simplify code.
Here are some examples:
Future<int> f1() async => 1;
Future<int> f1() => Future.value(1);
Future<void> f2() async {
throw Error();
}
Future<void> f2() {
return Future.error(Error());
}
According to official Dart Language tour, async executes synchronously until it finds await keyword :
Note: Although an async function might perform time-consuming operations, it doesn’t wait for those operations. Instead, the async function executes only until it encounters its first await expression (details). Then it returns a Future object, resuming execution only after the await expression completes.
Related
I am new to unit tests in Dart/Flutter and I would like to write a test for a void function. When functions return something I am writing a test like this:
test('Gets user save', () async {
final userSave = await mockSource!.getUserSave();
expect(userSave!.age, equals(20));
});
In such a scenario like above expect can be used since getUserSave function returns a user model.
How about checking if test passes of fails for a void/Future function like below? I can not use expect because it does not return a value.
Future<void> clearUserSave() async {
DatabaseClient mockDBClient = MockDatabaseClientImpl();
mockDBClient.clear();
}
I use flutter_test and mockito for testing.
Typically a void function will produce a side effect of some sort. When writing a test for a void function, I would check whatever state is effected by the void function before and after calling the function to be sure that the desired side effect has occurred.
In this specific case, you are calling clear on a DatabaseClient. I don't know the specifics of the DatabaseClient api, but I would construct a test where the client contains some data before calling clear, and then check that the data is no longer there after calling clear.
Something along the lines of this:
Future<void> clearUserSave() async {
DatabaseClient mockDBClient = MockDatabaseClientImpl();
mockDBClient.add(SOMEDATA);
expect(mockDBClient.hasData, true);
mockDBClient.clear();
expect(mockDBClient.hasData, false);
}
you can add nullable variable and assign variable value in method, after call method check if this varaible isNotNull
like this:
test('Gets user save', () async {
await mockSource?.getUserSave();
final userSave=mockSource.user;
expect(userSave,isNotNull );
});
testing a function that return void :
expect(
() async => await functionThatReturnsVoid(),
isA<void>(),
);
I have a function that calls a Future in it. Now I am not sure if the first function needs to be a future too to await for the data. Here is my code:
FireBaseHandler handler = FireBaseHandler();
saveRows() {
handler.saveRows(plan.planId, plan.rows); ///this is a future
}
In my FireBaseHandler class I have this Future:
final CollectionReference usersCol =
FirebaseFirestore.instance.collection('users');
Future saveRows(String id, data) async {
return await usersCol.doc(myUser.uid).collection('plans').doc(id)
.update({'rows': data});
}
So does the first function needs to be a Future too?
You can have async function inside a sync function. But this way you lose ability to await for the result. And await is allowed only in functions marked as async, which leads us to Future as a the result of that function. So, yes, if you need to wait for the result, both functions have to be Future functions.
Edit:
If you need your wrapper function to be sync, but still be able to retrieve the result from inner async function, you can use a callback:
saveRows(Function(dynamic) callback) {
handler.saveRows(plan.planId, plan.rows).then((result){
callback(result);
});
}
This way the result will be retrieved at any point of time after calling the function (the code is not awaited)
I'm trying to mock sharedPreferences using Mockito in my flutter project. Here is the error log.
package:mockito/src/mock.dart 190:7 Mock._noSuchMethod
package:mockito/src/mock.dart 184:45 Mock.noSuchMethod
test\feature\number_trivia\data\datasource\number_trivia_local_datasource_test.mocks.dart 67:14 MockSharedPreferences.setString
package:clean_arch_tdd/features/number_trivia/data/datasources/number_trivia_local_datasource.dart 31:30 NumberTriviaLocalDataSourceImpl.cacheLastNumberTrivia
test\feature\number_trivia\data\datasource\number_trivia_local_datasource_test.dart 51:18 main.<fn>.<fn>
MissingStubError: 'setString'
No stub was found which matches the arguments of this method call:
setString('CACHED_NUMBER_TRIVIA', '{"text":"Test trivia","number":1}')
Add a stub for this method using Mockito's 'when' API, or generate the mock for MockSharedPreferences with 'returnNullOnMissingStub: true'.
The error refer to this line of code.
local_data_source_test.dart
test('should call sharedPreferences to cache the data', () {
dataSource.cacheLastNumberTrivia(tNumberTriviaModel);
final expectedJsonString = jsonEncode(tNumberTriviaModel.toJson());
verify(mockSharedPreferences.setString(
cachedNumberTrivia, expectedJsonString));
});
local_data_source.dart
#override
Future<void> cacheLastNumberTrivia(NumberTriviaModel triviaToCache) {
return sharedPreferences.setString(
cachedNumberTrivia, jsonEncode(triviaToCache.toJson()));
}
It show that method setString from the mocked sharedPreferences is missing. I already run the pub command to generate the mocks. I also have some test case in the file that use the getString method. And it works fine.
Is there something I'm missing so I can't use the setString method? Or are there any solution to this problem?
Thx in advance.
I actually figured it out, sorry for not posting the answer immediately.
I found myself forgot to call the stub for the setString method. Here is the code.
group('cacheNumberTrivia', () {
const tNumberTriviaModel =
NumberTriviaModel(number: 1, text: 'Test trivia');
test('should call sharedPreferences to cache the data', () async {
when(mockSharedPreferences.setString(any, any))
.thenAnswer((_) async => true);
dataSource.cacheLastNumberTrivia(tNumberTriviaModel);
final expectedJsonString = jsonEncode(tNumberTriviaModel.toJson());
verify(mockSharedPreferences.setString(
cachedNumberTrivia, expectedJsonString));
});
});
The stub was actually this line:
when(mockSharedPreferences.setString(any, any))
.thenAnswer((_) async => true);
I forgot to call the when method before verifying the result with verify, hence it throws an error.
I have an async method. Can I call it from an non-async method? Like the following
My method
void method() async{
await 'something'
}
Case 1
onPressed:() {
method();
}
Case 2
onPressed:() async{
await method();
}
Which of the above is correct? It seems to me two of them is OK. However, the second one I think works much more slower, am I wrong?
In general, the caller of an async function must also be asynchronous if it wants to wait for the call to complete. This makes asynchronous-ness contagious.
In your case, your async function is a "fire-and-forget" function; callers cannot wait for it to complete, so it doesn't matter. Your second case (with await method()) is wrong because you should use await only on Future/FutureOr, but method returns void, so there nothing to wait for. (The Dart analyzer would warn you about this if you have the await_only_futures lint enabled.)
You also could simplify your code further by using a tear-off instead of creating an unnecessary closure:
onPressed: method
In Scala and other programming languages one can use Futures and Await.
(In real code one would use e.g. zip+map instead of Await)
def b1() = Future { 1 }
def b2() = Future { 2 }
def a() = Future {
Await.result(b1(),Duration.inf) + Await.result(b2(),Duration.inf)
}
What is the difference to Async/Await in Javascript/Scala?
async function b1() { return 1 }
async function b2() { return 3 }
async function a() {
return await b1() + await b2()
}
The "Await.result" function in Scala is "blocking", which means that the calling thread will be paused until the awaited Future is completed, at which point it will resume with the returned value.
Pausing a thread can be expensive in a system under high load, as the thread context has to be saved in memory, and it can cause cache misses etc.. Blocking threads is considered poor practice in concurrent programming for that reason.
The async / await syntax in Javascript is non-blocking. When an async function invokes "await", it is converted into a Future, and placed into the execution queue. When the awaited future is complete, the calling function is marked as ready for execution and it will be resumed at some later point. The important difference is that no Threads need to be paused in this model.
There are a number of libraries which implement the async / await syntax in Scala, including https://github.com/scala/scala-async
Further reading
The Futures and Promises book by PRASAD, PATIL, AND MILLER has a good introduction to blocking and non-blocking ops.
Await is blocking while async is non blocking. Await waits in the current thread to finish the task while async won't block the current thread and runs in background.
I don't have idea about JavaScript. In Scala Await from scala.concurrent package is used for blocking main thread. There's another library called scala-async what is used for using await inside async block.
If you are using scala-async, then you need to call await inside async