I have been trying to use dispatch_async in a method that returns a result. However, I observed that the method returns before executing the dispatch_async block. Due to this I'm not getting the results I expect. Here is some code that demonstrates my problem.
-(BOOL) isContactExists {
BOOL isContactExistsInXYZ = YES;
UserId *userId = contact.userId;
dispatch_async(dispatch_get_main_queue(), ^
{
iOSContact *contact = [iOSContact contactForUserId:userId];
if (nil == contact)
{
isContactExistsInXYZ = NO;
}
});
return isContactExistsInXYZ;
}
This method isContactExists is called somewhere else and based on the response from that method I have to do some stuff. But every time, the value of isContactExistsInXYZ is not what I expect. How do I handle dispatch_async in this situation?
If your going the block route your method needs to look something like this.
- (void)isContactExistsWithCompletionHandler:(void(^)(BOOL exists)) completion
{
dispatch_async(dispatch_get_main_queue(), ^
{
BOOL isContactExistsInXYZ = YES;
UserId *userId = contact.userId;
iOSContact *contact = [iOSContact contactForUserId:userId];
if (nil == contact)
{
isContactExistsInXYZ = NO;
}
completion(isContactExistsInXYZ);
});
}
And where you are calling it something like this.
[someObject isContactExistsWithCompletionHandler:^(BOOL exists) {
// do something with your BOOL
}];
You should also consider placing your heavy operations in a other que than main. Like this.
- (void)isContactExistsWithCompletionHandler:(void(^)(BOOL exists)) completion
{
dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, NULL);
dispatch_async(queue, ^
{
BOOL isContactExistsInXYZ = YES;
UserId *userId = contact.userId;
iOSContact *contact = [iOSContact contactForUserId:userId];
if (nil == contact)
{
isContactExistsInXYZ = NO;
}
dispatch_async(dispatch_get_main_queue(), ^
{
completion(isContactExistsInXYZ);
});
});
}
You need to respect that what you are trying to do is asynchronous and embrace that. This means not using a return value. Instead you can write your method to take a callback block as a parameter. Then, when your asynchronous check is complete you can call the block with the result.
So your method signature would become:
- (void)checkIfContactExistsWithCompletion:(ContactExistsBlock)completion {
Where ContactExistsBlock is a block definition with no return and probably a single BOOL parameter.
typedef void (^ContactExistsBlock) (BOOL exists);
The reason is dispatch_async(dispatch_get_main_queue(), ^does not wait until execution is done. You are probably messing up stuff there. Normally, this is used to update UI asynchronously along with other server content getting downloaded in some other thread. Try using dispatch_sync instead.
Related
I'm learning to develop iOS applications and now I'm reading some Objective-C source code.
This is a method to get user profile.
+ (void)getProfile:(void (^)(NSString *message))completion {
NSDictionary *dic = #{#"module":#"profile"};
[[self defaultManager] POST:KBaseUrl parameters:dic success:^(AFHTTPRequestOperation *operation, id responseObject) {
if ([self jsonOKForResponseObject:responseObject] && [self checkLogin:responseObject]) {
[ProfileManager sharedInstance].rank = responseObject[#"Variables"][#"space"][#"group"][#"grouptitle"];
[ProfileManager sharedInstance].credit = responseObject[#"Variables"][#"space"][#"credits"];
[ProfileManager sharedInstance].gender = responseObject[#"Variables"][#"space"][#"gender"];
completion(nil);
} else {
completion(#"fail");
}
} failure:^(AFHTTPRequestOperation *operation, NSError *error) {
completion(#"fail");
}];
}
My question is about the completion block.
I suppose that the completion block returns void and receives an NSString parameter.
In the block, what does completion(nil) mean?
Does that mean the block completion calls it self and send nil as parameter?
Doesn't that conflict with the parameter's type NSString*?
I'm not quite familiar with block in ObjC. Can anyone give a hint?
Yes you are right. It calls itself and sends nil as a parameter and it doesn't conflict with the NSString parameter. You are just passing nil to the NSString param.
You can call the above method like:
[YourClass getProfile:^(NSString *message) {
//message will be nill if you pass completion(nil);
}];
So when you pass the nill in the completion block, the message in the above method call will be nil!
The completion block is to notify you that your method call is complete, and at this point you can let that method pass certain paramteres , and if we consider your method:
+ (void)getProfile:(void (^)(NSString *message))completion {
NSDictionary *dic = #{#"module":#"profile"};
[[self defaultManager] POST:KBaseUrl parameters:dic success:^(AFHTTPRequestOperation *operation, id responseObject) {
if ([self jsonOKForResponseObject:responseObject] && [self checkLogin:responseObject]) {
.....
completion(#"hey sucess");
} else {
completion(#"if loop failed");
}
} failure:^(AFHTTPRequestOperation *operation, NSError *error) {
completion(#"some error occured");
}];
}
and when you call the method getProfile:
[YourClass getProfile:^(NSString *message) {
//Execution will reach here when the method call for getPRofile is complete and you have a result which you just sent through the completion block as a parameter which is a string and is waiting for you to process.
//you can do more from here
if([message isEqualToString:#"hey success"]){
//do something
}
if([message isEqualToString:#"if loop failed"]){
//do something
}
if([message isEqualToString:#"some error occured"]){
//do something
}
}];
As per #rmaddy comment, iis always a good practice to use BOOL to indicate the status success or fail rather than depending on a string as string can get localized/changed. We shold use the string to get more description of the error.
So your block should be:
+ (void)getProfile:(void (^)(BOOL status,NSString *message))completion {
.....
completion(YES,#"hey success");
}
and you can call it like":
[YourClass getProfile:^(BOOL status, NSString *message) {
if(status){
//get the success message
}else{
//get the fail message
}
}];
Blocks had a lot of type like :-
As a local variable:
returnType (^blockName)(parameterTypes) = ^returnType(parameters) {...};
As a property:
#property (nonatomic, copy, nullability) returnType (^blockName)(parameterTypes);
As a method parameter:
- (void)someMethodThatTakesABlock:(returnType (^nullability)(parameterTypes))blockName;
As an argument to a method call:
[someObject someMethodThatTakesABlock:^returnType (parameters) {...}];
As a typedef:
typedef returnType (^TypeName)(parameterTypes);
TypeName blockName = ^returnType(parameters) {...};
read more about it from here
Completion block does not return anything. It is just a piece of code to be executed, thats all. Though, you can give it some input where you call it so you can use the result elsewhere.
NSString *message is the input for your block so when you call your function getProfile as:
[MyClass getProfile:^(NSString *message) {
// write code to be executed when getProfile function finishes its job and sends message here.
}];
[MyClass getProfile:nil];
When used like this you're preferring not to do anything when getProfile finishes its job.
You are probably mixing your network manager's function's completion block with the one you wrote.
I have a senario that requires me to make multiple call to a web api. The following is an example.
getDataAsync:(NSDictionary *)dictionary withCompletion: (void (^)(NSDictionary*))completion {
__block int counter = n; // the number of async blocks
__block NSMutableDictionary *output = [[NSMutableDictionary alloc] init];
void (^returnBlock)(void) = ^{
counter--;
if(counter != 0) return;
completion(#{#"return": output});
return;
};
void (^getResourceA)(void) = ^{
[service getResourceA : dictionary[#"idA"] completion:
^(ServiceResult results, MyResourceA *a, NSString *errMsg) {
[output setValue:a.value forKey:a.name];
returnBlock();
}];
};
// followed by n-1 other blocks like getResourceA
//...
}
I want to use the built in dispatch_queue rather than my own custom solution here. How can I do that given the inner completion block used by the asynchronous service call?
Also any other advice on how to go about this would be appreciated.
Dispatch groups have been invented for this purpose:
dispatch_group_t requestGroup = dispatch_group_create();
dispatch_group_async(requestGroup, queue, ^{
// ...
});
dispatch_group_wait(requestGroup, DISPATCH_TIME_FOREVER);
completionBlock();
Or instead of waiting:
dispatch_group_notify(requestGroup, dispatch_get_main_queue(), ^{
completionBlock();
});
Also, instead of dispatching blocks to the group, you can also enter and leave a group manually, which works well with asynchronous service APIs:
dispatch_group_enter(requestGroup);
[service getResourceA : dictionary[#"idA"] completion: ^(ServiceResult results, MyResourceA *a, NSString *errMsg) {
[output setValue:a.value forKey:a.name];
dispatch_group_leave(requestGroup);
}];
Use dispatch_group_t. See Waiting on Groups of Queued Tasks.
The topic doesn't mention it, but use dispatch_group_notify to register a block instead of waiting inline.
I've working on/learning this all afternoon.
Following an example here: http://themainthread.com/blog/2012/09/communicating-with-blocks-in-objective-c.html, I have managed to setup a callback to get the result of an asynchronous call to a web service i have.
The web service takes a key code and the app transforms it and passes it back for authentication.
With my code below, how can I change the method from a void to an NSString that I can call to return my pass code?
-(void) showPassCode{
getAuthCodeAndMakePassCodeCompleteBlock callback = ^(BOOL wasSuccessful, NSString *passCode) {
if (wasSuccessful) {
NSLog(#"Code is: %#", passCode);
} else {
NSLog(#"Unable to fetch code. Try again.");
}
};
[self getAuthCodeAndMakePassCode:#"myAuthCode" withCallback:callback];
}
Ideally, I want it to work or look like this:
-(NSString *) strPassCode{
getAuthCodeAndMakePassCodeCompleteBlock callback = ^(BOOL wasSuccessful, NSString *passCode) {
if (wasSuccessful) {
return passCode;
} else {
return nil;
}
};
[self getAuthCodeAndMakePassCode:#"myAuthCode" withCallback:callback];
}
Without knowing the specifics of your code and how you query the server, I have to imagine it would look something like:
-(void)getAuthCodeWithCallback:(void (^)(NSString* authCode))callback
{
//make server call, synchronously in this example
NSString* codeReturnedFromServer = [self getServerCodeSynchronous];
callback(codeReturnedFromServer);
}
//some calling code
[self getAuthCodeWithCallback:^(NSString* authCode) {
NSLog(#"Code is: %#", authCode);
}];
If the method that gets your auth code from the server is asynchronous, it would look something like this:
-(void)getAuthCodeWithCallback:(void (^)(NSString* authCode))callback
{
//make server call, asynchronously in this example
[self someMethodCallToQueryCodeFromServerWithCallback:^(NSError* error, NSString* code) {
if (error) {
//handle error
}
else
callback(code);
}
}
I'd like to init a model, let the model do some async stuff and present a new viewcontroller once completed. But how do i wait for the two async methods to be completed and how do I setup the callback method?
Pseudocode
In my StartViewController.m:
-(void)openArticle
{
article = [Article initWithObject:someObject];
article.callback = changeView;
}
-(void)changeView
{
[self presentViewController:someController];
}
In my ArticleModel.m:
-(void)initWithObject:someObject
{
[self loadImage]
[self geoCode]
}
-(void)loadImage
{
runAsyncMethod: success:^() // This one is actually a AFNetworking setImageWithURLRequest
}
-(void)geoCode
{
runAnotherAsyncMethod: success:^() // This one is actually a geocodeAddressString operation
}
You can achieve this using dispatch_groups
- (void)initWithObject:(id)someObject
{
self = [super init];
if (self) {
self.dispatch_group = dispatch_group_create();
[self loadImage]
[self geoCode]
dispatch_group_notify(self.dispatch_group, dispatch_get_main_queue(), ^{
NSLog(#"Push new view controller");
});
}
return self;
}
- (void)loadImage
{
dispatch_group_enter(self.dispatch_group);
__weak __typeof(self) weakSelf = self;
runAsyncMethod: success:^{
__typeof(weakSelf) strongSelf = weakSelf;
if (strongSelf.dispatch_group) {
dispatch_group_leave(strongSelf.dispatch_group); // You need to ensure that this is called in both success and failure
}
}
}
- (void)geoCode
{
dispatch_group_enter(self.dispatch_group);
__weak __typeof(self) weakSelf = self;
runAnotherAsyncMethod: success:^{
__typeof(weakSelf) strongSelf = weakSelf;
if (strongSelf.dispatch_group) {
dispatch_group_leave(strongSelf.dispatch_group);
}
}
}
You do not wait. If you wait, it isn't asynchronous! You would be losing the entire point of asynchronous if you were to wait.
What you do is, when your success handler is called, you step out to the main thread (just in case you got called back on a background thread) and now do whatever you need to do. In other words, you just let your success handler get called whenever it happens to get called.
In your case, you might like to chain the things you want to do:
Call loadImage
In its callback, call geoCode
In its callback, step out to the main thread and present the new view controller.
You can use dispatch_group so that when a method is over, it just leaves the group. I use a similar code myself and it works like a charm.
- (void)initWithObject:someObject {
// Create a dispatch group
dispatch_group_t group = dispatch_group_create();
[self loadImageWithDispatchGroup:group];
[self geoCodeWithDispatchGroup:group];
// Here we wait for all the requests to finish
dispatch_group_notify(group, dispatch_get_main_queue(), ^{
// Do whatever you need to do when all requests are finished
});
}
- (void)loadImageWithDispatchGroup:(dispatch_group_t)group {
dispatch_group_enter(group);
runAsyncMethod: success:^() // This one is actually a AFNetworking setImageWithURLRequest
// In your success or failure AFNetworking method, call this as soon as the request ended
dispatch_group_leave(group);
}
- (void)geoCodeWithDispatchGroup:(dispatch_group_t)group {
dispatch_group_enter(group);
runAnotherAsyncMethod: success:^() // This one is actually a geocodeAddressString operation
// In your success async geocode callback method, call this as soon as the request ended
dispatch_group_leave(group);
}
I do not known your needs but native GCD way to wait several asynch tasks is
void dispatch_barrier_async(dispatch_queue_t queue, dispatch_block_t block);
https://developer.apple.com/library/ios/documentation/Performance/Reference/GCD_libdispatch_Ref/Reference/reference.html#//apple_ref/c/func/dispatch_barrier_async
i have a method, in which i want to accomplish a given task, however, the asynchronous commands and delegates made it difficult
i can do this :
- (void) fooPart1
{
...
SomeAssynchronousMethos * assync = [[SomeAssynchronousMethos alloc] init];
assync.delegate = self;
[assync start];
}
- (void) fooPart2
{
...
possibly some other assync
}
- (void)someAssynchronousMethosDelegateDidiFinish
{
[self fooPart2];
}
But isn't there a way to do smith. like this
- (void) foo
{
...
SomeAssynchronousMethos * assync = [[SomeAssynchronousMethos alloc] init];
assync.delegate = self;
[assync start];
wait for signal, but class is not blocked
...
possibly some other assync
}
- (void)someAssynchronousMethosDelegateDidiFinish
{
continue in foo after [assync start]
}
I don't like the idea of splitting a function to 2 or more parts, but is this the way how it is done in cocoa? or is there a better practice?
why i dont like this concept and searching for a better way of doing it :
lets say, i want to use a variable only for compleating a task - if i have everything in one function, i just use it, and than the var dies as i leave the function, if its split, i have to keep the var somehow around, until it doesnt finish
the code becomes fragmented and more difficult to read and maintain
may lead to bug
i end up with a set of part function, that needs to be called in precise order to accomplish one task (for which one function would be more suitable)
i used to make a thread and do only synchronous calls there, but not everything supports a synchronous call
what would be realy nice, is to have smth, like
- (void) foo
{
...
int smth = 5;
SomeAssynchronousMethos * assync = [[SomeAssynchronousMethos alloc] init];
assync.delegate = self;
#freez([assync start]);
// when freez - the local function variables are frozen
// (next commands in function are not excuted until unfreez)
// from outer look, it looks like the function finished
// however when unfreeze, it is called from this point on
//int smth is still 5
}
- (void)someAssynchronousMethosDelegateDidiFinish
{
#unfreez([assync start]);
}
when the execution would reach freez, it would store all local vars allocated in function and when called unfreez, it would continue from that point on, when it was freez with that command
This seems like an ideal application of a completion handler block.
Alter your start method to take a parameter which is a block and call it like so:
- (void) fooPart1
{
...
SomeAssynchronousMethos * assync = [[SomeAssynchronousMethos alloc] init];
[assync startOnComplete: ^(NSError* error) // example, you can have any params or none
{
// handle error if not nil
if (error != nil)
{
// do something with it
}
// code to do on completion
}];
}
Your start method would look something like this
-(void) startOnComplete: (void(^)(NSError*)) completionBlock
{
// copy the block somewhere
theSavedCompletionBlock = [completionBlock copy];
// kick off async operation
}
-(void) someMethodThatRunsAttheEndOfTheAsyncOp
{
theSavedCompletionBlock(nilOrError);
[theSavedCompletionBlock release];
}