I am trying to make this block.
+ (void)getUsersAndTracksWithSearch:(NSString *)search withCompletion:(void(^)(NSMutableArray *users, NSMutableArray *tracks)) completion
{
}
I am using SoundCloud API. I am using their API method for getting a list of tracks and users. which is:
[SCRequest performMethod:SCRequestMethodGET
onResource:[NSURL URLWithString:userURL]
usingParameters:nil
withAccount:nil
sendingProgressHandler:nil
responseHandler:^(NSURLResponse *response, NSData *data, NSError *error)
{
})];
I am using this method twice and getting two responses, so how should I set up a block so that on completion of the block I have both responses?
Thanks
Have a look at NSOperationQueue, it can be used in a way which chains their operations i.e. chains their blocks.
Related
Let's say I want to use the rest api of some service, Twitbookr, to get a user's profile information. So I get the user to log in via my app and I get the credentials needed to make the call.
After this I load a new ViewController. I want to populate the fields in this view with the user's profile information. So I make my first call to the api:
[NSURLConnection sendAsynchronousRequest:req
queue:NSOperation.mainQueue
completionHandler:
^(NSURLResponse *response, NSData *data, NSError *error)
{
// successfully fetch data using credentials
// assume complete code here including dispatch_async etc
self.data = data;
})];
Then I want to use the data:
self.nameLabel.text = [self.data objectForKey:#"name"];
self.userid = [self.data objectForKey:#"userid"];
But the request hasn't finished yet, so the program throws an error.
And I want to make more calls to get different types of data, for example, the pictures from the frontpage album, which rely on the userid which I can only get from the request above. So how do I make my next subsequent call making sure I already have the userid from the first call?
What's the correct way to handle this situation?
Should I be using synchronous requests instead?
Should I put up a loading symbol until all of my requests are done? If so, how do I test that the requests have actually finished? And what's the point of them being asynchronous?
Try this
[NSURLConnection sendAsynchronousRequest:req
queue:NSOperation.mainQueue
completionHandler:
^(NSURLResponse *response, NSData *data, NSError *error)
{
// successfully fetch data using credentials
self.data = data;
//Use data here
dispatch_async(dispatch_get_main_queue(), ^{
//Update UI here
});
})];
When you use AsynchronousRequest,use data in completion block. I think your code
self.nameLabel.text = self.data.name;
is wrong,because self.data looks like NSData,and it not has a property of name
Update:As Mike said,only update UI on main thread.
I need to get image information from server, such image name, image id. Then use image id as one of parameters to make post, get image actual data. More specific, there are three images I should get.
First, I use getImageInfo to get image information.
- (void)getImageInfo {
// compose request
NSUserDefaults *getUserInfo = [NSUserDefaults standardUserDefaults];
NSString *uid = [getUserInfo objectForKey:#"uid"];
NSString *checkCode = [getUserInfo objectForKey:#"checkCode"];
NSString *data = [NSString stringWithFormat:#"uid=%#&yangzhengma=%#", uid, checkCode];
NSURL *url = [NSURL URLWithString:#"http://121.199.35.173:8080/xihuan22dcloud/services/Shibietupianservice/serviceGetallshibietu"];
NSMutableURLRequest *request = [NSMutableURLRequest requestWithURL:url];
request.HTTPBody = [data dataUsingEncoding:NSUTF8StringEncoding];
request.HTTPMethod = #"POST";
[[self.session dataTaskWithRequest:request completionHandler:^(NSData *data, NSURLResponse *response, NSError *error){
if (!error) {
NSHTTPURLResponse *httpResp = (NSHTTPURLResponse*) response;
if (httpResp.statusCode == 200) {
// parse data in ram and put into images' imageInfos array
[self.images parseImageInfo:[[NSString alloc] initWithData:data encoding:NSUTF8StringEncoding]];
[self getImageRawData];
dispatch_async(dispatch_get_main_queue(), ^{
[self.tableView reloadData];
});
}
}
}] resume];}
Then I use getImageRawData to get three image data.
- (void)getImageRawData {
// compose request dynamically
NSUserDefaults *getUserInfo = [NSUserDefaults standardUserDefaults];
NSString *uid = [getUserInfo objectForKey:#"uid"];
NSString *checkCode = [getUserInfo objectForKey:#"checkCode"];
NSURL *url = [NSURL URLWithString:#"http://121.199.35.173:8080/xihuan22dcloud/services/Shibietupianservice/serviceGetthetupian"];
NSMutableURLRequest *request = [NSMutableURLRequest requestWithURL:url];
request.HTTPMethod = #"POST";
NSInteger count = 0;
for (ImageInformation *temp in self.images.imageInfos) {
NSString *data = [NSString stringWithFormat:#"uid=%#&yangzhengma=%#&tupianid=%#", uid, checkCode, temp.imageId];
request.HTTPBody = [data dataUsingEncoding:NSUTF8StringEncoding];[[self.session dataTaskWithRequest:request completionHandler:^(NSData *data, NSURLResponse *response, NSError *error){
// if client side is no errors, continue
if (!error) {
// if server side is no errors, continue
NSHTTPURLResponse *httpResp = (NSHTTPURLResponse*) response;
if (httpResp.statusCode == 200) {
NSLog(#"图片内容：%#", [[NSString alloc] initWithData:data encoding:NSUTF8StringEncoding]);
// in ram and put into images' imageRawData array
[self.images parseImageRawData:[[NSString alloc] initWithData:data encoding:NSUTF8StringEncoding] withImageId:temp.imageId withIndex:count];
// store data to disk
// NSString *path = [[NSString alloc] initWithFormat:#"image%#", temp.imageId];
// [FCFileManager writeFileAtPath:path content:data];
dispatch_async(dispatch_get_main_queue(), ^{
[self.tableView reloadData];
});
}
}
}] resume];
count++;
}}
Here, it will loop three times, three responses come back, only the last one is complete, the others carry a error message, or incomplete raw data sometimes. Now I'm diving into concurrency programming guide, I guess serial queue likely can solve this problem.
Output like this:
2014-12-16 22:38:48.739 WeddingNewVersion[997:83366] 图片内容：<ns:serviceGetthetupianResponse xmlns:ns="http://serviceimpl.my.com"><ns:return>error</ns:return></ns:serviceGetthetupianResponse>
2014-12-16 22:38:48.749 WeddingNewVersion[997:83366] 图片内容：<ns:serviceGetthetupianResponse xmlns:ns="http://serviceimpl.my.com"><ns:return>error</ns:return></ns:serviceGetthetupianResponse>
2014-12-16 22:38:51.943 WeddingNewVersion[997:83366] 图片内容：<ns:serviceGetthetupianResponse xmlns:ns="http://serviceimpl.my.com"><ns:return>/9j/...(complete data)...9k=%%226654474.0</ns:return></ns:serviceGetthetupianResponse>
parameters of requests:
2014-12-17 14:59:25.364 WeddingNewVersion[1875:226651] uid=6&yangzhengma=odWoDXWcBv1jOrEhywkq7L&tupianid=41
2014-12-17 14:59:25.368 WeddingNewVersion[1875:226651] uid=6&yangzhengma=odWoDXWcBv1jOrEhywkq7L&tupianid=42
2014-12-17 14:59:25.368 WeddingNewVersion[1875:226651] uid=6&yangzhengma=odWoDXWcBv1jOrEhywkq7L&tupianid=43
the problem is likely not in composing request.
------------------------------------------------update1-----------------------------------------------
I have tried to put data task of session into a serial queue. Disappointed, this is not working.
dispatch_async(self.serialQueue, ^{
[[self.session dataTaskWithRequest:request completionHandler:^(NSData *data, NSURLResponse *response, NSError *error){...}] resume];
});
Meanwhile, I make delegateQueue of session as nil, reference says if nil, the session creates a serial operation queue for performing all delegate method calls and completion handler calls.
Now I am still confused how to make it right.
-----------------------------------------------update2------------------------------------------------
I add [NSThread sleepForTimeInterval:0.5] into the block dispatched to serial queue.
dispatch_async(self.serialQueue, ^{
[[self.session dataTaskWithRequest:request completionHandler:^(NSData *data, NSURLResponse *response, NSError *error){...}] resume];
[NSThread sleepForTimeInterval:0.5];
});
It does not work. The three responses are complete, but they are all the same.
Thank you in advance!
I'm just guessing as I've never tried it, but possibly your data tasks are all using the same TCP port on your end.
That would be OK if they were serialized - one after the other, in sequence - but if they overlap, then the server would receive garbled HTTP requests:
GET /foo
GET /bar
GET /baz
What the server would see might be something like:
GET /fGET /baroo
GET /baz
That your third requests actually works OK might be an accident of the timing.
If you absolutely require the three requests to be issued simultaneously, there are ways to open three different ports on your end. I don't know how to do it with Cocoa and Objective-C, but you can certainly do it with C and Berkeley Socket system calls. The Cocoa / Cocoa Touch networking methods are just wrappers around sockets.
A couple of thoughts:
Your technique of using a single NSMutableURLRequest instance, and repeatedly mutating it for each request (while the prior requests are still in progress) is curious.
In the spirit of thread safety, I would use a separate NSMutableURLRequest for each concurrent request. You don't want to risk having your thread issuing these requests mutate the request object while some background thread performing one of the prior requests. (See Apple's Thread Safety Summary in the Threading Programming Guide in which they point out that mutable classes are not generally thread safe.)
Having said that, the NSURLConnection documentation leaves us with the impression that this request object would be copied, mitigating this problem. I don't see this sort of assurance in the NSURLSession documentation (though I suspect it does the same thing).
I don't think this is the problem here (if this was the problem, the problem would likely be more erratic than what you report, and besides, I suspect that NSURLSession is handling this gracefully, anyway), but as a matter of good thread-safe coding habits, it would be prudent to let each concurrent request have its own NSMutableURLRequest object.
You have confirmed that the information being used in the requests looks valid.
If you wanted to take this to the next level, you might use Charles (or Wire Shark or whatever tool you prefer) to observe the actual requests as they go out. These sorts of tools are invaluable for debugging these sorts of problems.
If you observe the requests in Charles and confirm that they are valid, then this categorically eliminates client-side issues from the situation.
What is curious is that you are not receiving NSError object from dataTaskWithRequest. Nor are you receiving statusCode other than 200 from your server. That means that your requests were successfully sent to the server and received by the server.
Instead, the server is processing the request, but is having a problem fulfilling the request. This leads me to wonder about the server code, itself. I suspect that there is something in the server code that is preventing concurrent operations from taking place (e.g., locking some shared resource, such as temp file or SQL table, for the duration of the request). I would take a hard look at the server code and make sure there are no potential contention issues.
Furthermore, I would modify the server code to not simply report "error", but rather to produce a meaningful error message (e.g. system provided error messages, error codes, etc.). Your server is detecting an error, so you should have it tell you precisely what that error was.
Note, I am explicitly not advising you to make your requests run sequentially. That is inadvisable. While it might solve the immediate problem, you pay a huge performance penalty doing that, and it's not scalable. And remember, you really must handle concurrent requests gracefully, as you're likely to have multiple users of the app at some point.
I would take a hard look at the server code, adding further debugging information to the error messages in order to track down the problem.
I put request into for loop, it works. The first thought of rob about NSMutableRequest and NSURLSession seems right, I'm trying to catch the whole idea. Thanks for rob's answer. Anyway, this is code.
for (ImageInformation *temp in self.images.imageInfos) {
// compose request dynamically
NSURL *url = [NSURL URLWithString:#"http://121.199.35.173:8080/xihuan22dcloud/services/Shibietupianservice/serviceGetthetupian"];
NSMutableURLRequest *request = [NSMutableURLRequest requestWithURL:url];
request.HTTPMethod = #"POST";
NSString *data = [NSString stringWithFormat:#"uid=%#&yangzhengma=%#&tupianid=%#", uid, checkCode, temp.imageId];
request.HTTPBody = [data dataUsingEncoding:NSUTF8StringEncoding];
// data task
dispatch_async(self.serialQueue, ^{
[[self.session dataTaskWithRequest:request completionHandler:^(NSData *data, NSURLResponse *response, NSError *error){
// if client side is no errors, continue
if (!error) {
// if server side is no errors, continue
NSHTTPURLResponse *httpResp = (NSHTTPURLResponse*) response;
if (httpResp.statusCode == 200) {
// in ram and put into images' imageRawData array
[self.images parseImageRawData:[[NSString alloc] initWithData:data encoding:NSUTF8StringEncoding] withImageId:temp.imageId];
// store data to disk
// [FCFileManager writeFileAtPath:path content:data];
// dispatch display image task to main
dispatch_async(dispatch_get_main_queue(), ^{
if ([self.images.imageDrawDatasDic count] == [self.images.imageInfos count]) {
[self.tableView reloadData];
}
});
}
}
}] resume];
[NSThread sleepForTimeInterval:0.5];
});
}
}
I have video on my server and i want to know in how much time my video requires to download from server in the device. So that i can calculate the download time and check if server is slow or not. Any help would be appreciated.
Here is my code:
NSURLRequest *request = [[NSURLRequest alloc] initWithURL:[NSURL URLWithString:url] cachePolicy:NSURLCacheStorageNotAllowed timeoutInterval:[[Configuration sharedInstance] fatchTimeOut]];
[GMURLConnection sendAsynchronousRequest:request queue:self.operationQueue type:type withSubType:subType completionHandler:^(NSURLResponse *response, NSData *data, NSError *error) {
//**********************Success Handling ********************
if (type == OperationTypeMeta) {
failCount = 0;
NSError *error = nil;
NSArray *serverResponse=[NSArray array];
if(data)
serverResponse = [NSJSONSerialization JSONObjectWithData:data options:NSJSONReadingMutableContainers error:&error];
}];
You cannot use sendAsynchronousRequest and monitor connection status. As per apple docs:
To retrieve the contents of a URL using a completion handler block: If
you do not need to monitor the status of a request, but merely need to
perform some operation when the data has been fully received, you can
call sendAsynchronousRequest:queue:completionHandler:, passing a block
to handle the results.
So you have to implement NSURLConnection delegate. If you like block style(as I really really do), just create new class which implement delegate methods and takes NSURLRequest, status block and completion block as params and call them in appropriate delegate methods. This way you will have block-styled asynchronous requests with status updates.
If you do not know how to measure remaining time - each time didReceiveData: method is called append data length to class variable, and calculate how much time passed since last method receive. This way you have velocity(bytes per second). Knowing how big is your downloaded file just divide remaining size to your velocity and you'll get remaining time. There are some neat algoritms out there to get neat/steady remaining time instead of just jumping from 30secs to 2 mins(if we assume our internet connection is tricky). I've personally used #Ben Dolman's solution posted over here:
averageSpeed = SMOOTHING_FACTOR * lastSpeed + (1-SMOOTHING_FACTOR) * averageSpeed;
With some tweaking and playing with SMOOTHING_FACTOR I've accomplished great results.
I suggest you to use AFHTTPRequestOperation. Like this:
AFHTTPRequestOperation* operation = [[AFHTTPRequestOperation alloc] initWithRequest:request];
[operation setDownloadProgressBlock:^(NSUInteger bytesRead, long long totalBytesRead, long long totalBytesExpectedToRead){
}];
You may use NSURLConnection with its delegate.
When download file first you get response from server which has total file size
- (void)connection:(NSURLConnection *)connection didReceiveResponse:(NSURLResponse *)response
{
NSHTTPURLResponse *httpResponse = (NSHTTPURLResponse*)response;
_expectedContentLength = [response expectedContentLength];
if (_expectedContentLength==NSURLResponseUnknownLength)
_expectedContentLength = 0;
}
After in - (void)connection:(NSURLConnection *)connection didReceiveData:(NSData *)data you may calculate downloading time from time spent to download portion of file and whole file size
I'm using synchronous requests for the first time and would love some help. (The code I'm writing is solely for my own use, and given its purposes synchronous requests are not a problem.)
The code gets data from a web page in a series, manipulates the data, moves on to the next page in the series, manipulates THAT data, and so on. I'm using a synchronous request because I need the connection to finish loading and the data to be manipulated before the function loops to the next page.
Here's my looping code:
-(NSData *)myMethod {
NSString *string;
NSData *data;
for (int x = 1; x<100; x++) {
string = [NSString stringWithFormat:#"http://www.blahblah.com/%d",(x)];
NSURLRequest *request = [NSURLRequest requestWithURL:[NSURL URLWithString:string]];
NSURLResponse *response = nil;
NSError *error = nil;
data = [NSURLConnection sendSynchronousRequest:request returningResponse:&response error:&error];
}
return data;
}
When I was using connectionWithRequest, I just put the code to manipulate the data in connectionDidFinishLoading and it worked fine. But with sendSynchronousRequest, even though NSLog shows that the loop code is looping, the code in connectionDidFinishLoading never runs.
How can I fix this?
(Or am I taking the wrong approach completely?)
Here's how to take #nhgrif's good advice to perform asynch and preserve all of the results.
- (void)doRequest:(NSInteger)requestIndex gatheringResultsIn:(NSMutableArray *)array completion:(void (^)(void))completion {
if (requestIndex < 100) {
NSString *string = [NSString stringWithFormat:#"http://www.blahblah.com/%d",(requestIndex)];
NSURLRequest *request = [NSURLRequest requestWithURL:[NSURL URLWithString:string]];
[NSURLConnection sendAsynchronousRequest:request queue:[NSOperationQueue mainQueue] completionHandler:^(NSURLResponse *response, NSData *data, NSError *error) {
if (data) [array addObject:data];
[self doRequest:requestIndex+1 gatheringResultsIn:array completion: completion];
}];
} else {
completion();
}
}
This will run 100 requests indexed 0..99 placing the results in a mutable array. Call it like this:
NSMutableArray *results = [NSMutableArray array];
[self doRequest:0 gatheringResultsIn:results completion:^{
NSLog(#"100 NSData objects should be here: %#", results);
}];
connectionDidFinishLoading is an NSURLConnection delegate method for when you've sent asynchronous requests. Normally, you'd implement this method to get the data that loaded, but you don't need to do this, as it's returned synchronously and assigned to your data variable.
I will note however, you are definitely taking a poor approach here.
First of all, if you'd use asynchronous requests here, you could query all 100 URLs as basically the same time and let them return in their own time.
But what's more problematic is what actually happens with your code.
We create a URL, send the synchronous request, and when it finishes, assign the return to data.
... then we loop. And do this 99 times. 99 times we make this synchronous request (to a different URL each time) and overwrite the data that the previous request loaded. And after the 100th time, we exit the loop and return the data we downloaded in the final request.
I have some difficulties to set up the correct configuration relative to sendAsynchronousRequest:queue:completionHandler: method (NSURLConnection class).
My scenario is the following:
I set up a singleton class that manages different NSURLConnections. This singleton istance has a NSOperation Queue (called downloadQueue) that makes a request to a web server and retrieves a string path (1).
Once done, the path is used to download a file within a web server (2). Finally, when the file has been correctly downloaded, I need to update the UI (3).
I figured out only the first request: the one through which I'm able to download the path. Could you suggest me a way to perform the other two steps?
Few questions here:
the download queue (downloadQueue) is not the main one, is it possible to open a new NSURLConnection in that queue? In other words, is it correct? (See comments in code snippets)
if the previous question is correct, how can I grab the main queue and update the UI?
Here the code snippet I use to perform the first step where downloadQueue is an instance variable that can be obtain through accessor mehods (#property/#synthesized);
// initializing the queue...
downloadQueue = [[NSOperation alloc] init];
// other code here...
[NSURLConnection sendAsynchronousRequest:urlRequest queue:[self downloadQueue] completionHandler:^(NSURLResponse *response, NSData *data, NSError *error) {
if([data length] > 0 && error == nil) {
// here the path (1)
// how to perform a second connection?
// what type of queue do I have to use?
}
}];
You're on the right track for performing your first download.
In the completion handler block after the first download, you're computing the URL that you'll need for a second download, right? Then you can perform that second download the same way: call +[NSURLConnection sendAsynchronousRequest:...] again with the new URL and the same queue. You can do this within the completion block for the first download.
To update the UI after the second download is done, switch to the main queue within the completion block for that download. You can do this with dispatch_async() or dispatch_sync() (in this case it doesn't matter which because you don't have further work to do on the download queue) and dispatch_get_main_queue(), or with -[NSOperationQueue addOperationWithBlock:] and +[NSOperationQueue mainQueue].
Your code should look something like this:
// init download queue
downloadQueue = [[NSOperationQueue alloc] init];
// (1) first download to determine URL for second
[NSURLConnection sendAsynchronousRequest:urlRequest queue:[self downloadQueue] completionHandler:^(NSURLResponse *response, NSData *data, NSError *error) {
if([data length] > 0 && error == nil) {
// set newURLRequest to something you get from the data, then...
// (2) second download
[NSURLConnection sendAsynchronousRequest:newURLRequest queue:[self downloadQueue] completionHandler:^(NSURLResponse *newResponse, NSData *newData, NSError *newError) {
if([newData length] > 0 && newError == nil) {
[[NSOperationQueue mainQueue] addOperationWithBlock:^{
// (3) update UI
}];
}
}];
}
}];
For updating the ui, as far as I know, you have to do that on the main thread. The ui could be updated from other threads but those updates are not fully reliable. In an app that I put together that made request to a web service, I make use of dispatch_async() to get access to the main queue and then I update, in my case a table view, from that call.
dispatch_async(dispatch_get_main_queue(), ^{
//block to be run on the main thread
[self.tableView reloadData];
});
I hope this helps.