I've noticed recently, that my code, that uses AFNetworking (latest version from master branch) stopped working properly under iOS 6.
Here's my code:
httpClient = [AFHTTPClient clientWithBaseURL:[NSURL URLWithString:baseURL]];
httpClient.operationQueue.maxConcurrentOperationCount = 1;
where httpClient is a class variable.
Next, I'm creating a request:
NSMutableURLRequest *signInRequest = [httpClient requestWithMethod:#"POST" path:#"/user/register" parameters:dataToSend];
signInRequest.timeoutInterval = 15.0;
signInRequest.cachePolicy = NSURLRequestReloadIgnoringLocalAndRemoteCacheData;
AFJSONRequestOperation *signInOperation = [AFJSONRequestOperation JSONRequestOperationWithRequest:signInRequest success:^(NSURLRequest *request, NSHTTPURLResponse *response, id JSON)
{
// Blah
}
failure:^(NSURLRequest *request, NSHTTPURLResponse *response, NSError *error, id JSON)
{
// Blah
}];
[httpClient enqueueHTTPRequestOperation:signInOperation];
All the other requests are constructed similarly. The first enqueued operation works well, I can get into success handler block. However, next calls to other requests are finished with fail handler and request timeout error, no matter how big is timeout value I choose.
I have done the same calls using plain NSURLConnection, writing a tons of code :), with success, requests were processed properly.
I switched to iOS 5 device, and the code above works fine.
I switched to 3G connection (iOS 6), and the code above works.
It seems like I have this problem only on WiFi connections (except the case when I'm in the same subnet with my REST server.)
Any thoughts on this?
Thank you in advance.
It seems you also posted an issue on AFNetworking's github and found the solution yourself :)!
It seems like iOS 6 changes something to TCP implementation or something. I moved server to third-party hosting and it is working now.
For future readers, the issue can be found here.
Related
I have been working on this problem for quite some time. I've read the many similar posts and their solutions don't help me (many posts are old and stale). I have a new Xamarin project and a localhost API that I am testing to. Initially I installed our production API locally and it doesn't connect. So I created a smaller test API and have had varying results in connecting. Xamarin Android will not connect with a "Failed to connect to localhost:XXXXX" message. I built a WPF project and get the same results. However I can connect to outside public API's (github) with both. I can connect to my localhost with a Windows forms desktop app and successfully get & post. Postman and httprepl both connect locally and successfully get & post. A couple of the other things the other posts suggested are 1) Client handler to ignore SSL - doesn't work, 2) Use the emulator's browser to connect to localhost - cannot connect. (this might be exposing where the problem really is but not sure).
Also The debugger seems to be behaving oddly. When the debugger gets to the HttpResponseMessageresponse.... line it rarely makes it to the next line so I cannot inspect the statuscode. The block I have below is in a try/catch and rarely makes it to the catch if the connection truly fails. Sometimes it will be caught at the catch and that's how I get the "failed to connect" message. Another developer here has run the project on his computer and gets the same results/errors on his computer.
HttpResponseMessage response = await client.SendAsync(request);
var Statuscode = response.StatusCode;
if (response.IsSuccessStatusCode)
{
var jsonbody = response.Content.ReadAsStringAsync().Result;
var ResponseBody = JsonConvert.DeserializeObject(response.Content.ReadAsStringAsync().Result);
}
else
{
string oops = "Nothing";
}
I seem to be running out of options and am looking for any advice.
Thanks in advance.
I'm having incredible difficulty setting up the Twilio Client in Meteor JS, and would really appreciate any help.
I have extracted the relevant code and error logs below. So far as I can tell, it should be simple. The code is just grabbing an authtoken which I have previously generated, and then trying to set up the device using that authtoken. But it's not working.
'click #initializeDevice'(event) {
var thisAuthToken = Session.get('myAuthToken');
console.log(thisAuthToken); // I have confirmed with Twilio support that these authtokens are correctly generated
const Device = require('twilio-client').Device;
Device.setup(thisAuthToken, { debug: true });
var myStatus = Device.status()
console.log(myStatus); //this is logging "offline"
Device.on('ready',function (device) {
log('Twilio.Device Ready!'); //this is not logging anything
});
},
When that code runs, it generates the following logs:
eyJhbGciDpvdXRnb2luZz9hcHBTaWQ9QVA2NDE2MzJmMzA1ZjJiY2I[Note:I have deleted part of the middle of the logged authtoken for the purpose of this public post]5YmMxOGQyOWVlNGU2ZGM0NjdmMzRiNDVhNCIsImV4cCI6MTU3Nz0ygbJKTx15GgNCWDkm-iUPjn_O1NZU6yovp4vjE
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 Setting up VSP
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 WSTransport.open() called...
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 Attempting to connect...
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 Closing and cleaning up WebSocket...
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 No WebSocket to clean up.
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 Could not connect to endpoint: ws does not work in the browser. Browser clients must use the native WebSocket object
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 Closing and cleaning up WebSocket...
modules.js?hash=69069bec9aeba9503ae3467590cf182be57d9e62:3605 No WebSocket to clean up.
calltemplate.js:31 offline
I'm doing this all from a local server, tunneled through NGROK. I've also set up the Twilio back end, linked the app, purchased a number, etc.
So far as I can tell, the issue, from the logs, appears to be something to do with the way that Meteor uses WebSockets.
Could not connect to endpoint: ws does not work in the browser. Browser clients must use the native WebSocket object
This is a not a Meteor related problem rather than browser issue.
Make sure your browser supports WebRTC
BTW, Your browser might be supporting it but you'd need to enable it.
I have a bunch of Xamarin iOS applications running on our Corporate iPhones and iPads. I recently upgraded my Visual Studio and Mac Book Pro to latest release version and my applications are no longer PROCESSING data from my web service correctly. I did not change version of Xamarin.Forms
VS 2017 Version 15.9.13
Xamarin.Forms.2.4.0.282
The web services are http. I have my NSAppTransportSecurityNSAllowsArbitraryLoads
set to allow http.
I am using the flag true to generate the Task.Factory.FromAsync methods in the Web Service Reference.cs file.
I could go on with describing how everything works if anyone needs more info.
Luckily, my Mac Book Pro Simulator has old and new versions of my apps installed in different iPhone/iPad devices for me to test. I installed fiddler to make sure all the requests/responses were the same. My old App in one device works and my new App in a different device fails. Same thing for ALL of my Applications. Old one works and new one fails.
When I review fiddler, the requests and responses are all working the same:
Request is made
App gets an authentication prompt
Authentication happens
Request returns code 200 with the data requested
The only difference in Fiddler is that two of the headers are in a different order: Working one has SOAPAction first and Content-Type next, the one failing has Content-Type and then SOAPAction. I am grasping at straws here because the order of these headers should not matter.
This is for all the applications. The only problem is that my new applications are ignoring the response. It is almost like the endMethod function from the parameter passed to FromAsync is never called.
This is the function call:
public async Task<ProcessLogon> ExecuteSaveLogon(DeviceItem DeviceItem)
{
ResultLogin = new ProcessLogon();
try
{
var ProcessLogon = await Task<WebServiceTrading.ProcessLogon>.Factory.FromAsync(serviceTrading.BeginSaveLogon, serviceTrading.EndSaveLogon, ToServiceTrading(DeviceItem), TaskCreationOptions.None);
ResultLogin = FromServiceTrading(ProcessLogon);
}
catch (Exception ex)
{
System.Diagnostics.Debug.WriteLine(#"ExecuteSaveLogon ERROR {0}", ex.Message);
}
return ResultLogin;
}
Does anyone know of any DEBUG or LOGGING tools that I could use to find out why my successful response is lost?
Last time we tested our application was on Tuesday, it was working well.
Today morning a long wait and timeout received from Microsoft when sending DETECT request with binary data in body. Nothing has changed between in our application.
I tested and my keys are okay (when they are not, we receive an answer but not 200, not a timeout).
When I try to DETECT using an URL, it is working (proper content-type and body of course)
Did anyone noticed the same, or using the API differently?
HttpClient client = HttpClientBuilder.create().build();
HttpPost request = new HttpPost("https://westeurope.api.cognitive.microsoft.com/face/v1.0/detect?returnFaceLandmarks=false&returnFaceId=true");
request.addHeader("Ocp-Apim-Subscription-Key", MY_KEY);
request.addHeader("Content-Type", "application/octet-stream");
HttpEntity entity = new ByteArrayEntity(Base64.getDecoder().decode(base64));
request.setEntity(entity);
response = client.execute(request);
It seems it was not a problem at out side. Suddenly it started to work...Hopefully will stay this way.
Stoping SignalR connection of Javascript client is slow (around 30 seconds).
I use SignalR, version 2.0.1 (but problem reproduced in 2.0.0 ) for webchat integrated to ASP.NET site. I have one .NET SignalR client(for other reason) and a lot of JS clients (chat clients). My test case is very simple. I want to disable chat on ASP.NET page. For this reason I try to stop SignalR connection using next code:
$.connection.hub.stop();
But onDisconnected method (on hub) was calling only after 30 seconds. I suppose that connection was stopped by disconnect timeout but not by Javascript code. I use LongPolling transpot by default.
var initObject = { transport: ["longPolling", "webSockets", "foreverFrame", "serverSentEvents"] };
$.connection.hub.start(initObject).done(function () {
...
}
Problem reproduce in last versions of Google Chrome but works fine in IE and Mozilla. What reason of so strange behaviour of SignalR and how can I avoid it?
For anyone else having this issue just use conn.Stop(new TimeSpan(0)) to kill the connection instantly.
conn.disconnect();
conn.stop();
hub.subscribe(null);
conn = null;
hub= null;
try this and let me know if it works.It works in android though.
The problem is solved by updating SignalR version to the 2.0.3.