I have an app that currently uses NSURLConnection for the vast majority of its networking. I would like to move to NSURLSession because Apple tells me that is the way to go.
My app just uses the synchronous version of NSURLConnection by way of the + (NSData *)sendSynchronousRequest:(NSURLRequest *)request returningResponse:(NSURLResponse **)response error:(NSError **)error class method. I do this within a NSBlockOperation running on an NSOperationQueue so I am not needlessly blocking the main queue. The big advantage to doing things this way is that I can make the operations dependent on one another. For example, I can have the task that is requesting data be dependent on the login task finishing.
I have not seen any support for synchronous operations within NSURLSession. All I can find are articles deriding me for even thinking of using it synchronously and that I am a horrible person for blocking the threads. Fine. But I see no way to make NSURLSessionTasks dependent on each other. Is there a way to do that?
Or is there a description of how I would do such a thing in a different way?
The harshest criticisms of synchronous network requests are reserved for those who do it from the main queue (as we know that one should never block the main queue). But you're doing it on your own background queue, which addresses the most egregious problem with synchronous requests. But you're losing some wonderful features that asynchronous techniques provide (e.g. cancelation of requests, if needed).
I'll answer your question (how to make NSURLSessionDataTask behave synchronously) below, but I'd really encourage you to embrace the asynchronous patterns rather than fighting them. I'd suggest refactoring your code to use asynchronous patterns. Specifically, if one task is dependent upon another, simply put the initiation of the dependent task in the completion handler of the prior task.
If you have problems in that conversion, then post another Stack Overflow question, showing us what you tried, and we can try to help you out.
If you want to make an asynchronous operation synchronous, a common pattern is to use a dispatch semaphore so your thread that initiated the asynchronous process can wait for a signal from the completion block of the asynchronous operation before continuing. Never do this from the main queue, but if you're doing this from some background queue, it can be a useful pattern.
You can create a semaphore with:
dispatch_semaphore_t semaphore = dispatch_semaphore_create(0);
You can then have the completion block of the asynchronous process signal the semaphore with:
dispatch_semaphore_signal(semaphore);
And you can then have the code outside of the completion block (but still on the background queue, not the main queue) wait for that signal:
dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER);
So, with NSURLSessionDataTask, putting that all together, that might look like:
[queue addOperationWithBlock:^{
dispatch_semaphore_t semaphore = dispatch_semaphore_create(0);
NSURLSession *session = [NSURLSession sharedSession]; // or create your own session with your own NSURLSessionConfiguration
NSURLSessionTask *task = [session dataTaskWithURL:url completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
if (data) {
// do whatever you want with the data here
} else {
NSLog(#"error = %#", error);
}
dispatch_semaphore_signal(semaphore);
}];
[task resume];
// but have the thread wait until the task is done
dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER);
// now carry on with other stuff contingent upon what you did above
]);
With NSURLConnection (now deprecated), you have to jump through some hoops to initiate requests from a background queue, but NSURLSession handles it gracefully.
Having said that, using block operations like this means that the operations won't respond to cancellation events (while they're running, at least). So I generally eschew this semaphore technique with block operations and just wrap the data tasks in asynchronous NSOperation subclass. Then you enjoy the benefits of operations, but you can make them cancelable, too. It's more work, but a much better pattern.
For example:
//
// DataTaskOperation.h
//
// Created by Robert Ryan on 12/12/15.
// Copyright © 2015 Robert Ryan. All rights reserved.
//
#import Foundation;
#import "AsynchronousOperation.h"
NS_ASSUME_NONNULL_BEGIN
#interface DataTaskOperation : AsynchronousOperation
/// Creates a operation that retrieves the contents of a URL based on the specified URL request object, and calls a handler upon completion.
///
/// #param request A NSURLRequest object that provides the URL, cache policy, request type, body data or body stream, and so on.
/// #param dataTaskCompletionHandler The completion handler to call when the load request is complete. This handler is executed on the delegate queue. This completion handler takes the following parameters:
///
/// #returns The new session data operation.
- (instancetype)initWithRequest:(NSURLRequest *)request dataTaskCompletionHandler:(void (^)(NSData * _Nullable data, NSURLResponse * _Nullable response, NSError * _Nullable error))dataTaskCompletionHandler;
/// Creates a operation that retrieves the contents of a URL based on the specified URL request object, and calls a handler upon completion.
///
/// #param url A NSURL object that provides the URL, cache policy, request type, body data or body stream, and so on.
/// #param dataTaskCompletionHandler The completion handler to call when the load request is complete. This handler is executed on the delegate queue. This completion handler takes the following parameters:
///
/// #returns The new session data operation.
- (instancetype)initWithURL:(NSURL *)url dataTaskCompletionHandler:(void (^)(NSData * _Nullable data, NSURLResponse * _Nullable response, NSError * _Nullable error))dataTaskCompletionHandler;
#end
NS_ASSUME_NONNULL_END
and
//
// DataTaskOperation.m
//
// Created by Robert Ryan on 12/12/15.
// Copyright © 2015 Robert Ryan. All rights reserved.
//
#import "DataTaskOperation.h"
#interface DataTaskOperation ()
#property (nonatomic, strong) NSURLRequest *request;
#property (nonatomic, weak) NSURLSessionTask *task;
#property (nonatomic, copy) void (^dataTaskCompletionHandler)(NSData * _Nullable data, NSURLResponse * _Nullable response, NSError * _Nullable error);
#end
#implementation DataTaskOperation
- (instancetype)initWithRequest:(NSURLRequest *)request dataTaskCompletionHandler:(void (^)(NSData * _Nullable data, NSURLResponse * _Nullable response, NSError * _Nullable error))dataTaskCompletionHandler {
self = [super init];
if (self) {
self.request = request;
self.dataTaskCompletionHandler = dataTaskCompletionHandler;
}
return self;
}
- (instancetype)initWithURL:(NSURL *)url dataTaskCompletionHandler:(void (^)(NSData * _Nullable data, NSURLResponse * _Nullable response, NSError * _Nullable error))dataTaskCompletionHandler {
NSURLRequest *request = [NSURLRequest requestWithURL:url];
return [self initWithRequest:request dataTaskCompletionHandler:dataTaskCompletionHandler];
}
- (void)main {
NSURLSessionTask *task = [[NSURLSession sharedSession] dataTaskWithRequest:self.request completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
self.dataTaskCompletionHandler(data, response, error);
[self completeOperation];
}];
[task resume];
self.task = task;
}
- (void)completeOperation {
self.dataTaskCompletionHandler = nil;
[super completeOperation];
}
- (void)cancel {
[self.task cancel];
[super cancel];
}
#end
Where:
//
// AsynchronousOperation.h
//
#import Foundation;
#interface AsynchronousOperation : NSOperation
/// Complete the asynchronous operation.
///
/// This also triggers the necessary KVO to support asynchronous operations.
- (void)completeOperation;
#end
And
//
// AsynchronousOperation.m
//
#import "AsynchronousOperation.h"
#interface AsynchronousOperation ()
#property (nonatomic, getter = isFinished, readwrite) BOOL finished;
#property (nonatomic, getter = isExecuting, readwrite) BOOL executing;
#end
#implementation AsynchronousOperation
#synthesize finished = _finished;
#synthesize executing = _executing;
- (instancetype)init {
self = [super init];
if (self) {
_finished = NO;
_executing = NO;
}
return self;
}
- (void)start {
if ([self isCancelled]) {
self.finished = YES;
return;
}
self.executing = YES;
[self main];
}
- (void)completeOperation {
self.executing = NO;
self.finished = YES;
}
#pragma mark - NSOperation methods
- (BOOL)isAsynchronous {
return YES;
}
- (BOOL)isExecuting {
#synchronized(self) {
return _executing;
}
}
- (BOOL)isFinished {
#synchronized(self) {
return _finished;
}
}
- (void)setExecuting:(BOOL)executing {
#synchronized(self) {
if (_executing != executing) {
[self willChangeValueForKey:#"isExecuting"];
_executing = executing;
[self didChangeValueForKey:#"isExecuting"];
}
}
}
- (void)setFinished:(BOOL)finished {
#synchronized(self) {
if (_finished != finished) {
[self willChangeValueForKey:#"isFinished"];
_finished = finished;
[self didChangeValueForKey:#"isFinished"];
}
}
}
#end
#Rob I would encourage you to post your reply as a solution, in view of the following documentation note from NSURLSession.dataTaskWithURL(_:completionHandler:):
This method is intended as an alternative to the
sendAsynchronousRequest:queue:completionHandler: method of
NSURLConnection, with the added ability to support custom
authentication and cancellation.
If semaphore based approach doesn't work, try polling based approach.
var reply = Data()
/// We need to make a session object.
/// This is key to make this work. This won't work with shared session.
let conf = URLSessionConfiguration.ephemeral
let sess = URLSession(configuration: conf)
let task = sess.dataTask(with: u) { data, _, _ in
reply = data ?? Data()
}
task.resume()
while task.state != .completed {
Thread.sleep(forTimeInterval: 0.1)
}
FileHandle.standardOutput.write(reply)
Polling based approach works very reliably, but effectively limits maximum throughput to polling interval. In this example, it's been limited to 10 times/sec.
I made a Swift package for this.
Semaphore based approach has been worked well so far, but since Xcode 11 era, it's getting broken. (maybe only for me?)
A data task does not finish if I wait for semaphores. If I wait for semaphore on different thread, it task fails with an error.
nw_connection_copy_protocol_metadata [C2] Client called nw_connection_copy_protocol_metadata on unconnected nw_connection error.
It seems something has been changed in the implementation as Apple is moving Network.framework.
Related
I have multiple GET API request methods that call a completion block when it finishes. Here is an example of one.
- (void)getUserInfo
onSuccess:(void (^)(id))successBlock
onFailure:(void (^)(NSError *))failureBlock {
NSString *urlStr = [NSString stringWithFormat:#"%#/user/", baseUrl];
[manager GET:urlStr parameters:nil progress:nil
success:^(NSURLSessionTask *task, id responseObject) {
successBlock(responseObject);
}
failure:^(NSURLSessionTask *operation, NSError *error) {
failureBlock(error);
}];
}
However, I noticed that I am repeating the manager GET request code in other methods. I want to create another method that handles all of the requests and remove the repetitive code. The URL seems to be the only thing that changes. However, there is one flaw. I need to call the successBlock to let the method know the request has finished.
Maybe I need to take another path altogether and do something different.
You can pass completion blocks around and then call them from your final method which handles all the get requests. I usually make completion blocks that are going to be reused typedefs for brevity. Here's an example of what I mean (I added a second example method that also passes through to the center getRequestWithURLString:onSuccess:onFailure: method):
LLFakeManager.h
#import <Foundation/Foundation.h>
NS_ASSUME_NONNULL_BEGIN
typedef void (^_Nullable SuccessCompletionBlock)(id responseObject);
typedef void (^_Nullable FailureCompletionBlock)(NSError *error);
#interface LLFakeManager : NSObject
- (void)getUserInfoOnSuccess:(SuccessCompletionBlock)successBlock onFailure:(FailureCompletionBlock)failureBlock;
- (void)getBooksCheckedOutOnSuccess:(SuccessCompletionBlock)successBlock onFailure:(FailureCompletionBlock)failureBlock;
#end
NS_ASSUME_NONNULL_END
LLFakeManager.m
#import "LLFakeManager.h"
#interface LLFakeManager()
- (void)getRequestWithURLString:(NSString *)urlString
onSuccess:(SuccessCompletionBlock)successBlock
onFailure:(FailureCompletionBlock)failureBlock;
#end
#implementation LLFakeManager
- (void)getUserInfoOnSuccess:(SuccessCompletionBlock)successBlock onFailure:(FailureCompletionBlock)failureBlock {
NSString *urlStr = #"FakeUserUrlPath";
[self getRequestWithURLString:urlStr onSuccess:successBlock onFailure:failureBlock];
}
- (void)getBooksCheckedOutOnSuccess:(SuccessCompletionBlock)successBlock onFailure:(FailureCompletionBlock)failureBlock {
NSString *urlString = #"FakeBooksUrlPath";
[self getRequestWithURLString:urlString onSuccess:successBlock onFailure:failureBlock];
}
// central method that will handle all the get requests
- (void)getRequestWithURLString:(NSString *)urlString
onSuccess:(SuccessCompletionBlock)successBlock
onFailure:(FailureCompletionBlock)failureBlock {
// some fake implementation here to do your request, then use the completion block passed in from whatever other method
if (successBlock) {
successBlock(#"responseObjectPassedBackHere");
}
}
#end
And an example of calling it:
LLFakeManager *manager = [[LLFakeManager alloc] init];
[manager getUserInfoOnSuccess:^(id _Nonnull responseObject) {
NSLog(#"Here's my response object = %#", responseObject);
} onFailure:^(NSError * _Nonnull error) {
// no implementation but same idea
}];
Would produce this log:
Here's my response object = responseObjectPassedBackHere
This site: http://goshdarnblocksyntax.com is a handy list of block syntax which may also be helpful for you.
The blocks -- if they have the same signature -- they can be passed along a chain of methods. Your GET blocks carry an unneccesary first parameter. The NSURLSessionTask *, if it is to be returned at all, should be returned synchronously. Moving that out of the block signature will allow you to standardize the blocks.
It's easier to say in code...
// changed this method name so it would compile
- (void)getUserInfoOnSuccess:(void (^)(id))successBlock
onFailure:(void (^)(NSError *))failureBlock {
NSString *urlStr = [NSString stringWithFormat:#"%#/user/", baseUrl];
// two things: get the task as a return value (if you need it)
// pass the blocks directly, without nesting them in new blocks
NSURLSessionTask *task = [manager GET: urlStr
parameters: nil
progress: nil
success: successBlock
failure: failureBlock];
// do something with the task
}
To make this work, alter the GET method return type and block signatures...
- (NSURLSessionTask *)GET:(NSString *)url parameters:(id)params progress:(id)progress success:(void (^)(id))successBlock failure:(void (^)(NSError *))failureBlock {
// return the session task created here
return task
}
As
NSURLConnection sendSynchronousRequest:returningResponse:error:&connectionError
is set deprecated I will have to replace an Importer I wrote a long time ago.
The Importer does the following:
It fetches data from API-A. The data there can be on multiple pages.
It uses the data from the first fetch (also multipage) to query data from API-B and merges
Results from the API-B query will be merged with data from API-A
I implemented this with a background-operation where I use methods for each API which called recursively if there are mulitple pages for the request.
But as NSURLSession does not support synchronous request I currently only see the option to have lot of overhead (e.g. iVars) control what's called in the completion block (e.g. next Page or start to query API-B).
So, what would be an elegant solution to bring this to NSURLSession.
NB: Just to make sure, my previous solution does not block the main thread at all. But back then it was the easiest way to control the merge of two sources.
This answer is not supposed to be best practice. It was just practical for me.
Faced with the situation when a bunch of synchronous requests is executed in background and the order of execution matters I've ended up using the following:
SyncRequestSender.h
#import <Foundation/Foundation.h>
#interface SyncRequestSender : NSObject
+ (NSData *)sendSynchronousRequest:(NSURLRequest *)request
returningResponse:(NSURLResponse **)response
error:(NSError **)error;
#end
SyncRequestSender.m
#import "SyncRequestSender.h"
#implementation SyncRequestSender
+ (NSData *)sendSynchronousRequest:(NSURLRequest *)request
returningResponse:(NSURLResponse **)response
error:(NSError **)error
{
dispatch_group_t group = dispatch_group_create();
dispatch_group_enter(group);
NSError __block *err = NULL;
NSData __block *data;
NSURLResponse __block *resp;
[[[NSURLSession sharedSession] dataTaskWithRequest:request
completionHandler:^(NSData* _data, NSURLResponse* _response, NSError* _error) {
resp = _response;
err = _error;
data = _data;
dispatch_group_leave(group);
}] resume];
dispatch_group_wait(group, DISPATCH_TIME_FOREVER);
if (response)
{
*response = resp;
}
if (error)
{
*error = err;
}
return data;
}
#end
This's an example in AFNetworking, it shows how to wait an asynchronous task.
- (NSArray *)tasksForKeyPath:(NSString *)keyPath {
__block NSArray *tasks = nil;
dispatch_semaphore_t semaphore = dispatch_semaphore_create(0);
[self.session getTasksWithCompletionHandler:^(NSArray *dataTasks, NSArray *uploadTasks, NSArray *downloadTasks) {
if ([keyPath isEqualToString:NSStringFromSelector(#selector(dataTasks))]) {
tasks = dataTasks;
} else if ([keyPath isEqualToString:NSStringFromSelector(#selector(uploadTasks))]) {
tasks = uploadTasks;
} else if ([keyPath isEqualToString:NSStringFromSelector(#selector(downloadTasks))]) {
tasks = downloadTasks;
} else if ([keyPath isEqualToString:NSStringFromSelector(#selector(tasks))]) {
tasks = [#[dataTasks, uploadTasks, downloadTasks] valueForKeyPath:#"#unionOfArrays.self"];
}
dispatch_semaphore_signal(semaphore);
}];
dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER);
return tasks;
}
I'm a bit confuse of how to take advantage of the new iOS 7 NSURLSession background transfers features and AFNetworking (versions 2 and 3).
I saw the WWDC 705 - What’s New in Foundation Networking session, and they demonstrated background download that continues after the app terminated or even crashes.
This is done using the new API application:handleEventsForBackgroundURLSession:completionHandler: and the fact that the session's delegate will eventually get the callbacks and can complete its task.
So I'm wondering how to use it with AFNetworking (if possible) to continue downloading in background.
The problem is, AFNetworking conveniently uses block based API to do all the requests, but if the app terminated or crashes those block are also gone. So how can I complete the task?
Or maybe I'm missing something here...
Let me explain what I mean:
For example my app is a photo messaging app, lets say that I have a PhotoMessage object that represent one message and this object has properties like
state - describe the state of the photo download.
resourcePath - the path to the final downloaded photo file.
So when I get a new message from the server, I create a new PhotoMessage object, and start downloading its photo resource.
PhotoMessage *newPhotoMsg = [[PhotoMessage alloc] initWithInfoFromServer:info];
newPhotoMsg.state = kStateDownloading;
self.photoDownloadTask = [[BGSessionManager sharedManager] downloadTaskWithRequest:request progress:nil destination:^NSURL *(NSURL *targetPath, NSURLResponse *response) {
NSURL *filePath = // some file url
return filePath;
} completionHandler:^(NSURLResponse *response, NSURL *filePath, NSError *error) {
if (!error) {
// update the PhotoMessage Object
newPhotoMsg.state = kStateDownloadFinished;
newPhotoMsg.resourcePath = filePath;
}
}];
[self.photoDownloadTask resume];
As you can see, I use the completion block to update that PhotoMessage object according to the response I get.
How can I accomplish that with a background transfer? This completion block won't be called and as a result, I can't update the newPhotoMsg.
A couple of thoughts:
You have to make sure you do the necessary coding outlined in the Handling iOS Background Activity section of the URL Loading System Programming Guide says:
If you are using NSURLSession in iOS, your app is automatically relaunched when a download completes. Your app’s application:handleEventsForBackgroundURLSession:completionHandler: app delegate method is responsible for recreating the appropriate session, storing a completion handler, and calling that handler when the session calls your session delegate’s URLSessionDidFinishEventsForBackgroundURLSession: method.
That guide shows some examples of what you can do. Frankly, I think the code samples discussed in the latter part of the WWDC 2013 video What’s New in Foundation Networking are even more clear.
The basic implementation of AFURLSessionManager will work in conjunction with background sessions if the app is merely suspended (you'll see your blocks called when the network tasks are done, assuming you've done the above). But as you guessed, any task-specific block parameters that are passed to the AFURLSessionManager method where you create the NSURLSessionTask for uploads and downloads are lost "if the app terminated or crashes."
For background uploads, this is an annoyance (as your task-level informational progress and completion blocks you specified when creating the task will not get called). But if you employ the session-level renditions (e.g. setTaskDidCompleteBlock and setTaskDidSendBodyDataBlock), that will get called properly (assuming you always set these blocks when you re-instantiate the session manager).
As it turns out, this issue of losing the blocks is actually more problematic for background downloads, but the solution there is very similar (do not use task-based block parameters, but rather use session-based blocks, such as setDownloadTaskDidFinishDownloadingBlock).
An alternative, you could stick with default (non-background) NSURLSession, but make sure your app requests a little time to finish the upload if the user leaves the app while the task is in progress. For example, before you create your NSURLSessionTask, you can create a UIBackgroundTaskIdentifier:
UIBackgroundTaskIdentifier __block taskId = [[UIApplication sharedApplication] beginBackgroundTaskWithExpirationHandler:^(void) {
// handle timeout gracefully if you can
[[UIApplication sharedApplication] endBackgroundTask:taskId];
taskId = UIBackgroundTaskInvalid;
}];
But make sure that the completion block of the network task correctly informs iOS that it is complete:
if (taskId != UIBackgroundTaskInvalid) {
[[UIApplication sharedApplication] endBackgroundTask:taskId];
taskId = UIBackgroundTaskInvalid;
}
This is not as powerful as a background NSURLSession (e.g., you have a limited amount of time available), but in some cases this can be useful.
Update:
I thought I'd add a practical example of how to do background downloads using AFNetworking.
First define your background manager.
//
// BackgroundSessionManager.h
//
// Created by Robert Ryan on 10/11/14.
// Copyright (c) 2014 Robert Ryan. All rights reserved.
//
#import "AFHTTPSessionManager.h"
#interface BackgroundSessionManager : AFHTTPSessionManager
+ (instancetype)sharedManager;
#property (nonatomic, copy) void (^savedCompletionHandler)(void);
#end
and
//
// BackgroundSessionManager.m
//
// Created by Robert Ryan on 10/11/14.
// Copyright (c) 2014 Robert Ryan. All rights reserved.
//
#import "BackgroundSessionManager.h"
static NSString * const kBackgroundSessionIdentifier = #"com.domain.backgroundsession";
#implementation BackgroundSessionManager
+ (instancetype)sharedManager {
static id sharedMyManager = nil;
static dispatch_once_t onceToken;
dispatch_once(&onceToken, ^{
sharedMyManager = [[self alloc] init];
});
return sharedMyManager;
}
- (instancetype)init {
NSURLSessionConfiguration *configuration = [NSURLSessionConfiguration backgroundSessionConfigurationWithIdentifier:kBackgroundSessionIdentifier];
self = [super initWithSessionConfiguration:configuration];
if (self) {
[self configureDownloadFinished]; // when download done, save file
[self configureBackgroundSessionFinished]; // when entire background session done, call completion handler
[self configureAuthentication]; // my server uses authentication, so let's handle that; if you don't use authentication challenges, you can remove this
}
return self;
}
- (void)configureDownloadFinished {
// just save the downloaded file to documents folder using filename from URL
[self setDownloadTaskDidFinishDownloadingBlock:^NSURL *(NSURLSession *session, NSURLSessionDownloadTask *downloadTask, NSURL *location) {
if ([downloadTask.response isKindOfClass:[NSHTTPURLResponse class]]) {
NSInteger statusCode = [(NSHTTPURLResponse *)downloadTask.response statusCode];
if (statusCode != 200) {
// handle error here, e.g.
NSLog(#"%# failed (statusCode = %ld)", [downloadTask.originalRequest.URL lastPathComponent], statusCode);
return nil;
}
}
NSString *filename = [downloadTask.originalRequest.URL lastPathComponent];
NSString *documentsPath = NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES)[0];
NSString *path = [documentsPath stringByAppendingPathComponent:filename];
return [NSURL fileURLWithPath:path];
}];
[self setTaskDidCompleteBlock:^(NSURLSession *session, NSURLSessionTask *task, NSError *error) {
if (error) {
// handle error here, e.g.,
NSLog(#"%#: %#", [task.originalRequest.URL lastPathComponent], error);
}
}];
}
- (void)configureBackgroundSessionFinished {
typeof(self) __weak weakSelf = self;
[self setDidFinishEventsForBackgroundURLSessionBlock:^(NSURLSession *session) {
if (weakSelf.savedCompletionHandler) {
weakSelf.savedCompletionHandler();
weakSelf.savedCompletionHandler = nil;
}
}];
}
- (void)configureAuthentication {
NSURLCredential *myCredential = [NSURLCredential credentialWithUser:#"userid" password:#"password" persistence:NSURLCredentialPersistenceForSession];
[self setTaskDidReceiveAuthenticationChallengeBlock:^NSURLSessionAuthChallengeDisposition(NSURLSession *session, NSURLSessionTask *task, NSURLAuthenticationChallenge *challenge, NSURLCredential *__autoreleasing *credential) {
if (challenge.previousFailureCount == 0) {
*credential = myCredential;
return NSURLSessionAuthChallengeUseCredential;
} else {
return NSURLSessionAuthChallengePerformDefaultHandling;
}
}];
}
#end
Make sure app delegate saves completion handler (instantiating the background session as necessary):
- (void)application:(UIApplication *)application handleEventsForBackgroundURLSession:(NSString *)identifier completionHandler:(void (^)())completionHandler {
NSAssert([[BackgroundSessionManager sharedManager].session.configuration.identifier isEqualToString:identifier], #"Identifiers didn't match");
[BackgroundSessionManager sharedManager].savedCompletionHandler = completionHandler;
}
Then start your downloads:
for (NSString *filename in filenames) {
NSURL *url = [baseURL URLByAppendingPathComponent:filename];
NSURLRequest *request = [NSURLRequest requestWithURL:url];
[[[BackgroundSessionManager sharedManager] downloadTaskWithRequest:request progress:nil destination:nil completionHandler:nil] resume];
}
Note, I don't supply any of those task related blocks, because those aren't reliable with background sessions. (Background downloads proceed even after the app is terminated and these blocks have long disappeared.) One must rely upon the session-level, easily recreated setDownloadTaskDidFinishDownloadingBlock only.
Clearly this is a simple example (only one background session object; just saving files to the docs folder using last component of URL as the filename; etc.), but hopefully it illustrates the pattern.
It shouldn't make any difference whether or not the callbacks are blocks or not. When you instantiate an AFURLSessionManager, make sure to instantiate it with NSURLSessionConfiguration backgroundSessionConfiguration:. Also, make sure to call the manager's setDidFinishEventsForBackgroundURLSessionBlock with your callback block - this is where you should write the code typically defined in NSURLSessionDelegate's method:
URLSessionDidFinishEventsForBackgroundURLSession:(NSURLSession *)session. This code should invoke your app delegate's background download completion handler.
One word of advice regarding background download tasks - even when running in the foreground, their timeouts are ignored, meaning you could get "stuck" on a download that's not responding. This is not documented anywhere and drove me crazy for some time. The first suspect was AFNetworking but even after calling NSURLSession directly, the behaviour remained the same.
Good luck!
AFURLSessionManager
AFURLSessionManager creates and manages an NSURLSession object based on a specified NSURLSessionConfiguration object, which conforms to <NSURLSessionTaskDelegate>, <NSURLSessionDataDelegate>, <NSURLSessionDownloadDelegate>, and <NSURLSessionDelegate>.
link to documentation here documentation
What are the best practices for making a serial queue of NSURLSessionTasks ?
In my case, I need to:
Fetch a URL inside a JSON file (NSURLSessionDataTask)
Download the file at that URL (NSURLSessionDownloadTask)
Here’s what I have so far:
session = [NSURLSession sharedSession];
//Download the JSON:
NSURLRequest *dataRequest = [NSURLRequest requestWithURL:url];
NSURLSessionDataTask *task =
[session dataTaskWithRequest:dataRequest
completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
//Figure out the URL of the file I want to download:
NSJSONSerialization *json = [NSJSONSerialization JSONObjectWithData:data options:kNilOptions error:nil];
NSURL *downloadURL = [NSURL urlWithString:[json objectForKey:#"download_url"]];
NSURLSessionDownloadTask *fileDownloadTask =
[session downloadTaskWithRequest:[NSURLRequest requestWithURL:[NSURL URLWithString:playlistURL]]
completionHandler:^(NSURL *location, NSURLResponse *response, NSError *error) {
NSLog(#"completed!");
}];
[fileDownloadTask resume];
}
];
Apart from the fact that writing a completion block within another completion looks messy, I am getting an EXC_BAD_ACCESS error when I call [fileDownloadTask resume]... Even though fileDownloadTask is not nil!
So, what is the best of way of sequencing NSURLSessionTasks?
You need to use this approach which is the most straight forward: https://stackoverflow.com/a/31386206/2308258
Or use an operation queue and make the tasks dependent on each others
=======================================================================
Regarding the HTTPMaximumConnectionsPerHost method
An easy way to implement a first-in first-out serial queue of NSURLSessionTasks is to run all tasks on a NSURLSession that has its HTTPMaximumConnectionsPerHost property set to 1
HTTPMaximumConnectionsPerHost only ensure that one shared connection will be used for the tasks of that session but it does not mean that they will be processed serially.
You can verify that on the network level using http://www.charlesproxy.com/, you wil discover that when setting HTTPMaximumConnectionsPerHost, your tasks will be still be started together at the same time by NSURLSession and not serially as believed.
Expriment 1:
Declaring a NSURLSession with HTTPMaximumConnectionsPerHost to 1
With task1: url = download.thinkbroadband.com/20MB.zip
With task2: url = download.thinkbroadband.com/20MB.zip
calling [task1 resume];
calling [task2 resume];
Result: task1 completionBlock is called then task2 completionBlock is called
The completion blocks might be called in the order you expected in case the tasks take the same amount of time however if you try to download two different thing using the same NSURLSession you will discover that NSURLSession does not have any underlying ordering of your calls but only completes whatever finishes first.
Expriment 2:
Declaring a NSURLSession with HTTPMaximumConnectionsPerHost to 1
task1: url = download.thinkbroadband.com/20MB.zip
task2: url = download.thinkbroadband.com/10MB.zip (smaller file)
calling [task1 resume];
calling [task2 resume];
Result: task2 completionBlock is called then task1 completionBlock is called
In conclusion you need to do the ordering yourself, NSURLSession does not have any logic about ordering requests it will just call the completionBlock of whatever finishes first even when setting the maximum number of connections per host to 1
PS: Sorry for the format of the post I do not have enough reputation to post screenshots.
Edit:
As mataejoon has pointed out, setting HTTPMaximumConnectionsPerHost to 1 will not guarantee that the connections are processed serially. Try a different approach (as in my original answer bellow) if you need a reliable serial queue of NSURLSessionTask.
An easy way to implement a first-in first-out serial queue of NSURLSessionTasks is to run all tasks on a NSURLSession that has its HTTPMaximumConnectionsPerHost property set to 1:
+ (NSURLSession *)session
{
static NSURLSession *session = nil;
static dispatch_once_t onceToken;
dispatch_once(&onceToken, ^{
NSURLSessionConfiguration *configuration = [NSURLSessionConfiguration defaultSessionConfiguration];
[configuration setHTTPMaximumConnectionsPerHost:1];
session = [NSURLSession sessionWithConfiguration:configuration];
});
return session;
}
then add tasks to it in the order you want.
NSURLSessionDataTask *sizeTask =
[[[self class] session] dataTaskWithURL:url
completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
#import "SessionTaskQueue.h"
#interface SessionTaskQueue ()
#property (nonatomic, strong) NSMutableArray * sessionTasks;
#property (nonatomic, strong) NSURLSessionTask * currentTask;
#end
#implementation SessionTaskQueue
- (instancetype)init {
self = [super init];
if (self) {
self.sessionTasks = [[NSMutableArray alloc] initWithCapacity:15];
}
return self;
}
- (void)addSessionTask:(NSURLSessionTask *)sessionTask {
[self.sessionTasks addObject:sessionTask];
[self resume];
}
// call in the completion block of the sessionTask
- (void)sessionTaskFinished:(NSURLSessionTask *)sessionTask {
self.currentTask = nil;
[self resume];
}
- (void)resume {
if (self.currentTask) {
return;
}
self.currentTask = [self.sessionTasks firstObject];
if (self.currentTask) {
[self.sessionTasks removeObjectAtIndex:0];
[self.currentTask resume];
}
}
#end
and use like this
__block __weak NSURLSessionTask * wsessionTask;
use_wself();
wsessionTask = [[CommonServices shared] doSomeStuffWithCompletion:^(NSError * _Nullable error) {
use_sself();
[self.sessionTaskQueue sessionTaskFinished:wsessionTask];
...
}];
[self.sessionTaskQueue addSessionTask:wsessionTask];
I use NSOperationQueue (as Owen has suggested). Put the NSURLSessionTasks in NSOperation subclasses and set any dependancies. Dependent tasks will wait until the task they are dependent on is completed before running but will not check the status (success or failure) so add some logic to control the process.
In my case, the first task checks if the user has a valid account and creates one if necessary. In the first task I update a NSUserDefault value to indicate the account is valid (or there is an error). The second task checks the NSUserDefault value and if all OK uses the user credentials to post some data to the server.
(Sticking the NSURLSessionTasks in separate NSOperation subclasses also made my code easier to navigate)
Add the NSOperation subclasses to the NSOperationQueue and set any dependencies:
NSOperationQueue *ftfQueue = [NSOperationQueue new];
FTFCreateAccount *createFTFAccount = [[FTFCreateAccount alloc]init];
[createFTFAccount setUserid:#"********"]; // Userid to be checked / created
[ftfQueue addOperation:createFTFAccount];
FTFPostRoute *postFTFRoute = [[FTFPostRoute alloc]init];
[postFTFRoute addDependency:createFTFAccount];
[ftfQueue addOperation:postFTFRoute];
In the first NSOperation subclass checks if account exists on server:
#implementation FTFCreateAccount
{
NSString *_accountCreationStatus;
}
- (void)main {
NSDate *startDate = [[NSDate alloc] init];
float timeElapsed;
NSString *ftfAccountStatusKey = #"ftfAccountStatus";
NSString *ftfAccountStatus = (NSString *)[[NSUserDefaults standardUserDefaults] objectForKey:ftfAccountStatusKey];
NSUserDefaults *userDefaults = [NSUserDefaults standardUserDefaults];
[userDefaults setValue:#"CHECKING" forKey:ftfAccountStatusKey];
// Setup and Run the NSURLSessionTask
[self createFTFAccount:[self userid]];
// Hold it here until the SessionTask completion handler updates the _accountCreationStatus
// Or the process takes too long (possible connection error)
while ((!_accountCreationStatus) && (timeElapsed < 5.0)) {
NSDate *currentDate = [[NSDate alloc] init];
timeElapsed = [currentDate timeIntervalSinceDate:startDate];
}
if ([_accountCreationStatus isEqualToString:#"CONNECTION PROBLEM"] || !_accountCreationStatus) [self cancel];
if ([self isCancelled]) {
NSLog(#"DEBUG FTFCreateAccount Cancelled" );
NSUserDefaults *userDefaults = [NSUserDefaults standardUserDefaults];
[userDefaults setValue:#"ERROR" forKey:ftfAccountStatusKey];
}
}
In the next NSOperation post data:
#implementation FTFPostRoute
{
NSString *_routePostStatus;
}
- (void)main {
NSDate *startDate = [[NSDate alloc] init];
float timeElapsed;
NSString *ftfAccountStatusKey = #"ftfAccountStatus";
NSString *ftfAccountStatus = (NSString *)[[NSUserDefaults standardUserDefaults] objectForKey:ftfAccountStatusKey];
if ([ftfAccountStatus isEqualToString:#"ERROR"])
{
// There was a ERROR in creating / accessing the user account. Cancel the post
[self cancel];
} else
{
// Call method to setup and run json post
// Hold it here until a reply comes back from the operation
while ((!_routePostStatus) && (timeElapsed < 3)) {
NSDate *currentDate = [[NSDate alloc] init];
timeElapsed = [currentDate timeIntervalSinceDate:startDate];
NSLog(#"FTFPostRoute time elapsed: %f", timeElapsed);
}
}
if ([self isCancelled]) {
NSLog(#"FTFPostRoute operation cancelled");
}
}
This question already has answers here:
Problem understanding Delegating pattern & callback functions in Objective C
(3 answers)
Closed 9 years ago.
I work on iOS platform , I want to know what is a delegate function and what is a callback function ? what is the difference between the two types of function or they are the same ??
example of delegate function is numberOfRowsInSection in UITableViewDelegate protocol and example of callback function is didReceiveLocalNotification in appDelegate.m
Can we create our own callback function in Objective-C , if YES ,give an example ...
Thank you..
A couple of thoughts:
You suggest that didReceiveLocationNotification was a "callback function", but it's actually just a delegate method of the UIApplicationDelegate protocol. So, both numberOfRowsInSection and didReceiveLocalNotification are simply delegate methods.
Something more akin to a generic callback function would be the selector when scheduling a NSTimer or defining the handler for a UIGestureRecognizer, where the choice of method name is not predetermined.
Or callbacks are used extensively in CFArray.
But, the root of your question is less on the terminology, but rather a question of how to define an interface where the caller can specify a method that some other object will invoke (asynchronously) at some future date. There are a couple of common patterns:
Block parameter to method: It is increasingly common to define methods that take a block as a parameter. For example, you can have a method that is defined as follows:
- (NSURLSessionTask *)downloadAsynchronously:(NSURL *)url filename:(NSString *)filename completion:(void (^)(NSData *results, NSString *filename))completion {
NSURLSessionTask *task = [[NSURLSession sharedSession] dataTaskWithURL:url completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
dispatch_async(dispatch_get_main_queue(), ^{
completion(data, filename);
});
}];
[task resume];
return task;
}
That third parameter, completion, is a block of code that will be called with the download is done. Thus, you can invoke that method as follows:
[self downloadAsynchronously:url filename:filename completion:^(NSData *results, NSString *filename) {
NSLog(#"Downloaded %d bytes", [results length]);
[results writeToFile:filename atomically:YES];
}];
NSLog(#"%s done", __FUNCTION__);
You'll see that "done" message appear immediately, and that completion block will be called when the download is done. It admittedly takes a while to get used to the ungainly mess of punctuation that constitutes a block variable/parameter definition, but once you're familiar with the block syntax, you'll really appreciate this pattern. It eliminates the disconnect between the invoking of some method and the defining of some separate callback function.
If you want to simplify the syntax of dealing with blocks as parameters, you can actually define a typedef for your completion block:
typedef void (^DownloadCompletionBlock)(NSData *results, NSString *filename);
And then the method declaration, itself, is simplified:
- (NSURLSessionTask *)downloadAsynchronously:(NSURL *)url filename:(NSString *)filename completion:(DownloadCompletionBlock)completion {
NSURLSessionTask *task = [[NSURLSession sharedSession] dataTaskWithURL:url completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
dispatch_async(dispatch_get_main_queue(), ^{
completion(data, filename);
});
}];
[task resume];
return task;
}
Delegate-protocol pattern: The other common technique for communicating between objects is the delegate-protocol pattern. First, you define the protocol (the nature of the "callback" interface):
#protocol DownloadDelegate <NSObject>
- (NSURLSessionTask *)didFinishedDownload:(NSData *)data filename:(NSString *)filename;
#end
Then, you define your class that will be invoking this DownloadDelegate method:
#interface Downloader : NSObject
#property (nonatomic, weak) id<DownloadDelegate> delegate;
- (instancetype)initWithDelegate:(id<DownloadDelegate>)delegate;
- (NSURLSessionTask *)downloadAsynchronously:(NSURL *)url filename:(NSString *)filename;
#end
#implementation Downloader
- (instancetype)initWithDelegate:(id<DownloadDelegate>)delegate {
self = [super init];
if (self) {
_delegate = delegate;
}
return self;
}
- (NSURLSessionTask *)downloadAsynchronously:(NSURL *)url filename:(NSString *)filename {
NSURLSessionTask *task = [[NSURLSession sharedSession] dataTaskWithURL:url completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
dispatch_async(dispatch_get_main_queue(), ^{
[self.delegate didFinishedDownload:data filename:filename];
});
}];
[task resume];
return task;
}
#end
And finally, the original view controller which uses this new Downloader class must conform to the DownloadDelegate protocol:
#interface ViewController () <DownloadDelegate>
#end
And define the protocol method:
- (void)didFinishedDownload:(NSData *)data filename:(NSString *)filename {
NSLog(#"Downloaded %d bytes", [data length]);
[data writeToFile:filename atomically:YES];
}
And perform the download:
Downloader *downloader = [[Downloader alloc] initWithDelegate:self];
[downloader downloadAsynchronously:url filename:filename];
NSLog(#"%s done", __FUNCTION__);
Selector pattern: A pattern that you see in some Cocoa objects (e.g. NSTimer, UIPanGestureRecognizer) is the notion of passing a selector as a parameter. For example, we could have defined our downloader method as follows:
- (NSURLSessionTask *)downloadAsynchronously:(NSURL *)url filename:(NSString *)filename target:(id)target selector:(SEL)selector {
id __weak weakTarget = target; // so that the dispatch_async won't retain the selector
NSURLSessionTask *task = [[NSURLSession sharedSession] dataTaskWithURL:url completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
dispatch_async(dispatch_get_main_queue(), ^{
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Warc-performSelector-leaks"
[weakTarget performSelector:selector
withObject:data
withObject:filename];
#pragma clang diagnostic pop
});
}];
[task resume];
return task;
}
You'd then invoke that as follows:
[self downloadAsynchronously:url
filename:filename
target:self
selector:#selector(didFinishedDownload:filename:)];
But you also have to define that separate method that will be called when the download is done:
- (void)didFinishedDownload:(NSData *)data filename:(NSString *)filename {
NSLog(#"Downloaded %d bytes", [data length]);
[data writeToFile:filename atomically:YES];
}
Personally, I find this pattern to be far too fragile and dependent upon coordinating interfaces without any assistance from the compiler. But I include it for a bit of historical reference given that this pattern is used quite a bit in Cocoa's older classes.
Notifications: The other mechanism to provide the results of some asynchronous method is to send a local notification. This is generally most useful when either (a) the potential recipient of the results of the network request is unknown at the time the request was initiated; or (b) there may be multiple classes that want to be informed of this event. So the network request can post a notification of a particular name when it's done, and any object that is interested in being informed of this event can add themselves as an observer for that local notification through the NSNotificationCenter.
This is not a "callback" per se, but does represent another pattern for an object to be informed of the completion of some asynchronous task.
Those are a few examples of "callback" patterns. Clearly, the example provided was arbitrary and trivial, but hopefully it should give you an idea of your alternatives. The two most common techniques, nowadays, are blocks and delegate patterns. Blocks are increasingly being preferred when a simple and elegant interface is needed. But for rich and complicated interfaces, delegates are very common.