We Keep Coding

Making a method async in Xamarin (C#)

I have this method that downloads data from the internet:
public ObservableCollection<Magnetka> ParseJSON() {
ObservableCollection<Models.Magnetka> Markers = new ObservableCollection<Models.Magnetka>();
HttpWebRequest request = (HttpWebRequest)WebRequest.Create("(URL)");
request.Headers.Add("Authentication-Token", "(KEY)");
request.Method = "GET";
request.ContentType = "application/json";
CookieContainer cookieContainer = new CookieContainer();
request.CookieContainer = cookieContainer;
var response = request.GetResponse();
var responseString = new StreamReader(response.GetResponseStream()).ReadToEnd();
......... JSON parsing)
return Markers;
Now I found out this isn't too fast and it freezes the app for two seconds or so, so I need to make it async. However, I call the method like this in another class:
GetMarkers getMarkers = new GetMarkers();
Markers = getMarkers.ParseJSON();
How can I make my method async? I don't understand the concept well and if I make ParseJSON() async Task, I don't know where to put the "await" keyword.
Can anyone help me?

first, mark the method async and use the async version of GetResponse
public async Task<ObservableCollection<Magnetka>> ParseJSON()
{
...
var response = await request.GetResponseAsync();
...
return Markers;
}
then when you call it use the await keyword
Markers = await getMarkers.ParseJSON();

You would change your method to something like this instead:
public async Task<ObservableCollection<Magnetka>> ParseJSON()
{
var markers = new ObservableCollection<Magnetka>();
HttpWebRequest request = (HttpWebRequest)WebRequest.Create("(URL)");
request.Headers.Add("Authentication-Token", "(KEY)");
request.Method = "GET";
request.ContentType = "application/json";
CookieContainer cookieContainer = new CookieContainer();
request.CookieContainer = cookieContainer;
using var response = await request.GetResponseAsync().ConfigureAwait(false);
using var responseStream = response.GetResponseStream();
using var streamReader = new StreamReader(responseStream);
var json = await streamReader.ReadToEndAsync().ConfigureAwait(false);
// json parsing
return markers;
}
Firstly, you would mark your method async. This signals the compiler to set up a state machine for this method for continuation. Nothing async happening yet though.
Another thing you need to do is to wrap your return value in a Task, which is similar to a promise in some other languages. Tasks are await-able, which is where the asynchronous stuff is happening.
Switching from using the synchronous methods in WebRequest to use the Async variants and awaiting their results makes your method async.
So from:
request.GetResponse()
To:
await request.GetResponseAsync()
Which returns a awaitable Task<WebResponse>.
The same goes for the stream reader code, you can use its async API too.
A tricky thing by using asynchronous code is that, it kind of forces you to do this all the way up. However, there are exceptions for lifecycle methods and events, where you don't necessarily have to change the signature to return a Task.
Another thing you might notice here. I've added ConfigureAwait(false) at the end of the async method calls. I've done this because out of the box, after returning from an awaited method, the code will try to return to the originating context. This can lead to a couple of issues such as dead-locks if the originating context is the UI Thread. Or, it can lead to bad performance trying to switch between context a lot. Adding ConfigureAwait(false) will just return on the same context the await was run on, which is fine as long as we don't need or expect to return on that context.
As for where you call ParseJson() is probably equally important. Consider deferring it to some lifecycle method or event where you are not doing much else.
This could be OnResume() on Android or ViewWillAppear on iOS. If you are using MVVM, use a Command to encapsulate it.

Issue with async Task<T> in xamarin forms

I'm trying to use the Azure media services with REST api, in a xamarin shared project on visual studio 2013. This is the code i use to get the access token.
public HttpFactory()
{
string token = GetToken(serviceURI).Result;
//some more methods also async tasks
}
private async Task<string> GetToken(Uri serviceUri)
{
HttpWebRequest request = (HttpWebRequest)HttpWebRequest.Create(serviceURI);
request.Accept = "application/json";
request.Method = "GET";
request.Headers["Host"] = "media.windows.net";
request.Headers["x-ms-version"] = "2.9";
request.Headers["Authorization"] = "Bearer " + token;
var response = (HttpWebResponse) await request.GetResponseAsync();
if (response.StatusCode == HttpStatusCode.MovedPermanently)
{
serviceURI = new Uri(response.Headers["Location"]);
HttpWebRequest req = ( HttpWebRequest)WebRequest.Create(serviceURI);
var res = (HttpWebResponse)await req.GetResponseAsync();
using (Stream responseStream = res.GetResponseStream())
{
using (StreamReader reader = new StreamReader(responseStream))
{
string str = reader.ReadToEnd();
// var test = JsonConvert.DeserializeObject(str);
JToken jtoken = JsonConvert.DeserializeObject<JToken>(str);
return jtoken["access_token"].Value<string>();
}
}
}
return "";
}
But when the compiler reaches -
var response = (HttpWebResponse) await request.GetResponseAsync();
it skips the rest of the code, and i never get the response. I know the code is working - because it works just fine without the task, in a async void method.
Anyone knows how to fix this, or am i doing something wrong? I also tried this in vs2015 but its the same.

You have a deadlock over the UI thread.
You're blocking the thread with Task.Result when it is needed to complete the async method which will complete the task that it's waiting on.
That's why you shouldn't block synchronously on asynchronous code. You should await the task returned from GetToken instead:
string token = await GetToken(serviceURI);
If you can't use async in that method, either move that logic to a different method (e.g. OnLoad event handler).
Another solution would be to use ConfigureAwait on the GetResponseAsync task and so the rest of the method wouldn't run on the UI thread, avoiding the deadlock:
var response = (HttpWebResponse) await request.GetResponseAsync().ConfigureAwait(false);

ASP.NET WEB API do i need async client

I have a Web API application which exposes certain async methods, such as:
public async Task<HttpResponseMessage> Put(string id){
var data = await Request.Content.ReadAsStringAsync();
//do something to put data in db
return Request.CreateResponse(HttpStatusCode.Ok);
}
I also have a class library that consumes this API with System.Net.WebRequest like this:
var request = (HttpWebRequest)WebRequest.Create(url);
var response = await request.GetResponseAsync();
Does the response need to be retrievend async? If so, why? Or can I also use request.GetResponse();.
I used the GetResponse before which would sometimes throw a 500 error ( An asynchronous module or handler completed while an asynchronous operation was still pending). Once I changed it to GetResponseAsync(), I stopped receiving this error.
EDIT: Code that sometimes throws a 500 error
I had the web api method stripped down to (just to check whether the 500 error was business logic or something else). Note: this is after changing the consumer to async (first function is the consumer, then the api method):
public HttpWebResponse(GetApiResponseForPut(string url, string putData, NameValueCollection headers)
{
var request = (HttpWebRequest)WebRequest.Create(url);
request.CookieContainer = _cookieContainer;
request.Method = "PUT";
if (headers != null)
{
request.Headers.Add(headers);
}
var encoding = new ASCIIEncoding();
var byte1 = new byte[0];
if (putData != null)
{
byte1 = encoding.GetBytes(putData);
}
request.ContentType = "application/x-www-form-urlencoded";
request.ContentLength = byte1.Length;
Stream requestStream = request.GetRequestStream();
requestStream.Write(byte1, 0, byte1.Length);
requestStream.Close();
var response = await request.GetResponseAsync();
return (HttpWebResponse)response;
}
public async Task<HttpResponseMessage> Put(string id)
{
HttpResponseMessage response = Request.CreateResponse(HttpStatusCode.OK);
return response;
}

Does the response need to be retrievend async? If so, why?
No, it doesn't. You need to separate how the server-side operates and how the client queries your endpoint. When you expose a method which is async Task, you state that in your server-side calls, you're making async calls. This is transparent to the caller, all he gets is an API endpoint, he doesn't have any knowledge of your internal implementation.
Remember, even if you use async on the server-side, the request will only return to the caller once complete.

Cancel an async webrequest?

I am using the Async CTP library for Windows Phone. Does anyone know how to cancel a pending webrequest?
Request = (HttpWebRequest)WebRequest.Create(url);
Request.Credentials = new NetworkCredential(_settings.Username, _settings.Password);
WebResponse resp;
try
{
resp = await Request.GetResponseAsync();
}
There is no cancellation token (as specified in the ASYNC Ctp tap document).

You could try calling Request.Abort().

Is there a faster way to check if an external web page exists?

I wrote this method to check if a page exists or not:
protected bool PageExists(string url)
{
try
{
Uri u = new Uri(url);
WebRequest w = WebRequest.Create(u);
w.Method = WebRequestMethods.Http.Head;
using (StreamReader s = new StreamReader(w.GetResponse().GetResponseStream()))
{
return (s.ReadToEnd().Length >= 0);
}
}
catch
{
return false;
}
}
I am using it to check a set of pages (iterates from AAAA-AAAZ), and it takes between 3 and 7 seconds to run the entire loop. Is there a faster or more efficient way to do this?

I think your approach is rather good, but would change it into only downloading the headers by adding w.Method = WebRequestMethods.Http.Head; before calling GetResponse.
This could do it:
HttpWebRequest request = (HttpWebRequest)WebRequest.Create("http://www.example.com");
request.Method = WebRequestMethods.Http.Head;
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
bool pageExists = response.StatusCode == HttpStatusCode.OK;
You may probably want to check for other status codes as well.

static bool GetCheck(string address)
{
try
{
HttpWebRequest request = WebRequest.Create(address) as HttpWebRequest;
request.Method = "GET";
request.CachePolicy = new RequestCachePolicy(RequestCacheLevel.NoCacheNoStore);
var response = request.GetResponse();
return (response.Headers.Count > 0);
}
catch
{
return false;
}
}
static bool HeadCheck(string address)
{
try
{
HttpWebRequest request = WebRequest.Create(address) as HttpWebRequest;
request.Method = "HEAD";
request.CachePolicy = new RequestCachePolicy(RequestCacheLevel.NoCacheNoStore);
var response = request.GetResponse();
return (response.Headers.Count > 0);
}
catch
{
return false;
}
}
Beware, certain pages (eg. WCF .svc files) may not return anything from a head request. I know because I'm working around this right now.
EDIT - I know there are better ways to check the return data than counting headers, but this is a copy/paste from stuff where this is important to us.

One obvious speedup is to run several requests in parallel - most of the time will be spent on IO, so spawning 10 threads to each check a page will complete the whole iteration around 10 times faster.

You could do it using asynchronous way, because now you are waiting for results after each request. For few pages, you could just throw your function in ThreadPool, and wait for all requests to finish. For more requests, you could use asynchronous methods for your ResponseStream() (BeginRead etc.).
The other thing that can help you (help me for sure) is to clear .Proxy property:
w.Proxy = null;
Without this, at least 1st request is much slower, at least on my machine.
3. You can not download whole page, but download only header, by setting .Method to "HEAD".

I simply used Fredrik Mörk answer above but placed it within a method:
private bool checkURL(string url)
{
bool pageExists = false;
try
{
HttpWebRequest request = (HttpWebRequest)WebRequest.Create(url);
request.Method = WebRequestMethods.Http.Head;
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
pageExists = response.StatusCode == HttpStatusCode.OK;
}
catch (Exception e)
{
//Do what ever you want when its no working...
//Response.Write( e.ToString());
}
return pageExists;
}