We Keep Coding

Universal Windows 10 WebView - how to clear/disable cache?

I am having a lot of trouble clearing the WebView cache in my UWP app.
If I edit the content of a JS file linked from my HTML page, I can't get the change into my app unless I re-install the app.
The static WebView.ClearTemporaryWebDataAsync() method doesn't seem to work.
I have also tried adding headers to the request to disable caching:
private void reloadPage()
{
string url = getUrl();
HttpRequestMessage request = new HttpRequestMessage(HttpMethod.Get, new Uri(url));
request.Headers.Add("Cache-Control", "no-cache, no-store, must-revalidate");
request.Headers.Add("Pragma", "no-cache");
myWebView.NavigateWithHttpRequestMessage(request);
}
I also tried the following, on a punt (I'm not sure if this affects the WebView's caching behaviour), but still no joy:
private void onWebviewLoaded(object sender, RoutedEventArgs e)
{
Windows.Web.Http.Filters.HttpBaseProtocolFilter myFilter = new Windows.Web.Http.Filters.HttpBaseProtocolFilter();
myFilter.CacheControl.WriteBehavior = Windows.Web.Http.Filters.HttpCacheWriteBehavior.NoCache;
myFilter.CacheControl.ReadBehavior = Windows.Web.Http.Filters.HttpCacheReadBehavior.Default;
WebView.ClearTemporaryWebDataAsync().AsTask().Wait();
reloadPage();
}
Any help would be very much appreciated!
EDIT (14/12/15):
I have found that adding headers to the request (as in the first code example above) does work, but only if this has been in place for the lifetime of the app, since install. Which makes sense, as it's just saying not to cache this particular request - it could still use an old cached version.
This works as a cludge for now, but it would be much nicer to be able to make use of caching (e.g. for the duration of an app session), then later clear the cache (e.g. on next startup).
EDIT (14/07/16): The above approach doesn't seem to bear out. Caching behaviour seems to be erratic in the webview...
On a clean install of the app, I can see changes to CSS/JS files with absolutely NO code to clear/disable cache. Then at some seemingly arbitrary point, files seem to be cached, and I cannot clear them from the cache.

This finally works for me:
await WebView.ClearTemporaryWebDataAsync();
Windows.Web.Http.Filters.HttpBaseProtocolFilter myFilter = new Windows.Web.Http.Filters.HttpBaseProtocolFilter();
var cookieManager = myFilter.CookieManager;
HttpCookieCollection myCookieJar = cookieManager.GetCookies(new Uri("http://www.msftncsi.com/ncsi.txt"));
foreach (HttpCookie cookie in myCookieJar)
{
cookieManager.DeleteCookie(cookie);
}

I don't know if it helps but try to add a timestamp behind the url (url + "?=" + sometimestamphere)

Refreshing the webview seems to do a reload:
mainWebView.Refresh();
It's definitely a hack, but you may be able to insert at that some point in your application lifecycle to force reloading your content. Perhaps after the "mainWebView_NavigationCompleted()" event?

I had a similar problem with css in an UWP app not getting cleared. I ran ProcessMon and found that the UWP app was caching .css and .js files in Windows 10 at: C:\Users\\AppData\Local\Packages\microsoft.windows.authhost.sso.p_8wekyb3d8bbwe\AC\INetCache\Q8IHZDMV. I'm using the Web Authentication Broker so the scenario might be slightly different. If a location similar to this doesn't pan-out, you might want to try to run processmon (part of the SysInternals suite) when you start-up your UWP app to see if you can identify the path the UWP app is using for caching these assets. Once there, deleting these assets and restarting the app did the trick.

Facebook multi account

I want to write a simple win-forms tool for managing different social accounts.I have a problem if user has several Facebook accounts. When I try to log in as second user I just got redirecting page in webBrowser.I tried to use InternetSetOption and the INTERNET_SUPPRESS_COOKIE_PERSIST flag but it seems it does not help. How can I resolve this problem?
Login code
var lParameters = new Dictionary<string, object>();
lParameters["client_id"] = AppId;
lParameters["redirect_uri"] = "https://www.facebook.com/connect/login_success.html";
lParameters["response_type"] = "token";
lParameters["display"] = "popup";
lParameters["scope"] = "user_about_me";
Uri lUri = mFacebookClient.GetLoginUrl(lParameters);
Main.webBrowser.Navigate(lUri);
ADDED:
Maybe I'm doing something wrong with InternetSetOption? Sorry for newbie question. I really should use it like in this answer How do I use InternetSetOption? It looks difficult...

Most probable cause is because Facebook is setting a cookie with the session ID. INTERNET_SUPPRESS_COOKIE_PERSIST only makes cookies non-persistent, they will be cleared after the browser is destroyed or browser session finished, so if you are using the same instance they will still be alive.
You can finish your browsing session with InternetSetOption(0, 42, NULL, 0); (taken from here: http://social.msdn.microsoft.com/Forums/eu/csharpgeneral/thread/76a38eee-3ef6-4993-a54d-3fecc4eb6cff), so combining INTERNET_SUPPRESS_COOKIE_PERSIST and INTERNET_OPTION_END_BROWSER_SESSION (is what the 42 means) it should be cleared and ready for a new login.

Cookie persistence and best practice

In our application there are two types of user, let's call them Alpha and Beta users. Each of these users sees a different type of toolbar / menu.
We have decided to track this using cookies. The majority of our pages are either Alpha pages or Beta pages and then there are some common pages that Alpha and Beta users share. So in each view of our application where we know the user type (Alpha or Beta) we have the following code:
HttpCookie isAlphaCookie = new HttpCookie("IsAlpha", "false"); // or true
HttpCookie isBetaCookie = new HttpCookie("IsBeta", "true"); // or false
isAlphaCookie.Expires = DateTime.MaxValue;
isBetaCookie.Expires = DateTime.MaxValue;
Response.Cookies.Add(isAlphaCookie);
Response.Cookies.Add(isBetaCookie);
The idea is then, in common pages, we don't set any cookie and rely on the previously set cookie to determine which toolbar to load. So, these two cookies are set to true or false as above in our known pages before we read them in the controller method which loads our toolbar like so:
HeaderViewModel header = new HeaderViewModel
{
FirstName = UserProfile.CurrentUser.FirstName,
LastName = UserProfile.CurrentUser.LastName,
ImageUrl = null,
OrganisationName = UserProfile.CurrentUser.OrganisationName,
OrganisationUrl = UserProfile.CurrentUser.OrganisationUrl,
ShowAlphaToolbar = bool.Parse(Request.Cookies["IsAlpha"].Value),
ShowBetaToolbar = bool.Parse(Request.Cookies["IsBeta"].Value),
ShowPublicToolbar = false
};
return PartialView("Common/_Header", header);
From reading up on how to read / write cookies this seems to be the right approach; writing the cookie to the Response object and reading the cookie from the Request object.
The problem I'm having is that when i get to the controller method that loads the toolbar the values of the IsAlpha and IsBeta cookies are both empty strings and this breaks the application.
I have confirmed that the cookies are set in the Response before they are read in the Request.
I'm wondering whether I'm missing something fundamental here.

I only expect your assumption " in common pages, we don't set any cookie and rely on the previously set cookie to determine which toolbar to load" to work if you are calling these partial actions, which you referring to as "common pages" I guess, through Ajax Calls. If you are using
#Html.RenderAction('nameOfaAtion')
then I don't think what you have in place will work.
Reason is Both your main action and partial actions are executed within the same Http Request Cycle so the cookies you are trying to access from the Request Object in your common pages have not yet came as part of Request.
Edit
As I can see you are hard coding the cookies on each page so Guess you can also do something like below. Not originally the way you are trying to do but I think it will do the same thing you are trying.
In your pages change the partial view call to like this.
#Html.Action("LoadHeader", "Profile", new{IsAlpha=False, IsBeta=true});
Then Change the signature of the LoadHeader action to receive these two extra parameters
public ViewAction LoadHeader(bool IsAlpha, Bool IsBeta)
Then with the viewModel Initialization change two lines as below.
ShowAlphaToolbar = IsAlpha,
ShowBetaToolbar = IsBeta,

HttpWebRequest: Add Cookie to CookieContainer -> ArgumentException (Parametername: cookie.Domain)

I'm trying to login to a website via my application.
What I did:
First I figured out how the browser does the authorization process with Fiddler.
I examined how the POST request is built and I tried to reconstruct it.
The browser sends 4 cookies (Google Analytics) and I tried to set them:
CookieContainer gaCookies = new CookieContainer();
gaCookies.Add(new Cookie("__utma", "#########.###########.##########.##########.##########.#"));
gaCookies.Add(new Cookie("__utmb", "#########.#.##.##########"));
gaCookies.Add(new Cookie("__utmc", "#########"));
gaCookies.Add(new Cookie("__utmz", "#########.##########.#.#.utmcsr=(direct)|utmccn=(direct)|utmcmd=(none)"));
(just replaced the original cookie data with #)
Then I went through the code with the debugger and as soon as the first gaCookies.Add is executed, the application stops with an
System.ArgumentException: The parameter '{0}' cannot be an empty string. Parameter name: cookie.Domain
I would like to know why this happens. The constructor of Cookie doesn't require a domain and I don't know where I can get this value?
Would be very great if someone of you could help me with this.
I'm not a webdeveloper or an expert in web stuff so I don't know much about it.
Is there maybe a great source where I can learn about this if there is no "short and quick answer"?

CookieContainers can hold multiple cookies for different websites, therefor a label (the Domain) has to be provided to bind each cookie to each website. The Domain can be set when instantiating the individual cookies like so:
Cookie chocolateChip = new Cookie("CookieName", "CookieValue") { Domain = "DomainName" };
An easy way to grab the domain to is to make a Uri (if you aren't using one already) that contains your target url and set the cookie's domain using the Uri.Host property.
CookieContainer gaCookies = new CookieContainer();
Uri target = new Uri("http://www.google.com/");
gaCookies.Add(new Cookie("__utmc", "#########") { Domain = target.Host });

Grabbing Cookies in Web Browser Control - WP7

In order to log into a certain part of a website the users of my application require their cookie. To do this I need to grab it and pass it to url.
Does anyone know how to grab a certain websites cookie from the browser control?
I saw this method but wasn't quite clear.
Thanks, TP.

As of WP 7.1 Mango "release", if one may call it, please see the WebBrowser Control Overview for Windows Phone. It has been recently updated a little bit, and it turns out that they actually have added some support for cookie-retrieval from the WebBrowser. On the bottom of the page you will find a tiny link GetCookies(WebBrowser) pointing to description of a new class: WebBrowserExtensions with this very handy method. Yes, this class has only that one single member. It's an extension method, I suppose no explanations needed on that.
I have not played with this method much, but it seems that this will allow you to access the very same thing as the JS trick: the cookieset for the current URL. It probably will not allow to set anything, nor to peek cookies for other URLs. Maybe if you play hard with the CookieContainer you will receive, but I doubt.
On the 7.0 release, I've been struggling quite hard to achieve "cookie transparency" for my App. Long story short, my app was doing some background HTTP requests, and also had a WebBrowser to show some online content -- and "it would be great" if both sources of connections would emit the same cookies to the server.. And guess what, my application had to make the first request, then let the browser navigate. With such requirements, there was virtually is no way to achieve consistency of the cookies - bah, even with the current new and glorious GetCookie method, I suppose it would be damn hard. So, to the point - it was possible, but needed to use some hidden API, that is present publicitly on the Phone, but is hidden in the SDK. The API is available the (public) class System.Net.Browser.WebRequestCreator, freely available. The quirk is: in the SDK this class has a single public static property "IWebRequestCreate ClientHttp" with a method "Create" that you can use to "factory" your "raw http" connections - in case you dont want to use the WebClient for some reason. On the phone, and on the emulator, there is a second public static property called "IWebRequestCreate BrowserHttp", easily returned by Reflection:
PropertyInfo brwhttp = typeof(System.Net.Browser.WebRequestCreator)
.GetProperty("BrowserHttp")
with this property, you will be able to obtain a "special" internal instance of IWebRequestCreate that is used internally by the WebBrowser. By opening your background HTTP requests with this class, you will get your cookies automatically set as if they were created/sent by the WebBrowser control, but in turn - you will NOT be able to modify http headers, userprovide http user authentication and neither do a few lowlevel things - because all that settings will be synced with the WebBrowser's data stored for current 'system user instance', if I'm allowed to call it as such on the single-user Phone device heh. The interoperation between connections and the WebBrowser works both ways - if your HTTP connection (created with use of the 'hidden property') receives any settings/cookies/etc -- then the WebBrowser will instantly notice them and update its own cache. No cookie/session loss on neither of the sides!
If you need to passively get cookies for your subsequent connections after some first WebBrowser navigation - please use the GetCookie or the JS way.
But if you need your code to be first, and then pass authz to the WebBrowser -- you will probably have to dig deeper and use the above.. It's been hidden, so please resort to the other means first!
..and don't ask me how did I found it or how much time it took :P
have a nice fun with it
//edit: I've just found out, that the BrowserHttp property is a normal Silverlight's way to access the Browser's connection factory, please see BrowserHttp. It seems that it is only has been hidden in the 'miniSilverlight' written for the WP7 platform!

The approach being described in the post you linked is to use the WebBrowser control's InvokeScript method to run some javascript. However the post appears to use a "cookies" collection which doesn't actually exist.
string cookie = myWebBrowser.InvokeScript("document.cookie") as string;
Now for the hard part the string you get contains all pertinent cookie name/value pairs for the page with the values still being Url encoded. You will need to parse the returned string for the value you need.
See document.cookie property documentation.
Edit:
Looking at it fresh instead of relying on the post, InvokeScript invokes named function on the window of the host browser. Hence the page being displayed in the WebBrowser would itself need to include a function like:-
function getCookie() { return document.cookie; }
Then the InvokeScript would look like:-
string cookie = myWebBrowser.InvokeScript("getCookie");

As #quetzalcoatl already suggested, you can use internal instance of WebRequestCreator to share cookies between browser instances and instances of WebRequest. You don't get to access the cookies directly though, I think that's just a security measure by Microsoft.
This code below creates a WebReqeust object, connected to CookieContainer of WebBrowser instance. It then posts to a url to log in the user and store cookies in the container.
After it's done, all browser instances within the app instance will have required set of cookies.
var browser = new WebBrowser();
var brwhttp = typeof (WebRequestCreator).GetProperty("BrowserHttp");
var requestFactory = brwhttp.GetValue(browser, null) as IWebRequestCreate;
var uri = new Uri("https://www.login.com/login-handler");
var req = requestFactory.Create(uri);
req.Method = "POST";
var postParams = new Dictionary<string, string> {
{"username", "turtlepower"},
{"password": "ZoMgPaSSw0Rd1"}
};
req.BeginGetRequestStream(aReq => {
var webRequest = (HttpWebRequest)aReq.AsyncState;
using (var postStream = webRequest.EndGetRequestStream(aReq)) {
// Build your POST request here
var postDataBuilder = new StringBuilder();
foreach (var pair in paramsDict) {
if (postDataBuilder.Length != 0) {
postDataBuilder.Append("&");
}
postDataBuilder.AppendFormat("{0}={1}", pair.Key, HttpUtility.UrlEncode(pair.Value));
}
var bytes = Encoding.UTF8.GetBytes(postDataBuilder.ToString());
postStream.Write(bytes, 0, bytes.Length);
}
// Receive response
webRequest.BeginGetResponse(aResp => {
var webRequest2 = (HttpWebRequest) aResp.AsyncState;
webRequest = (HttpWebRequest)aResp.AsyncState;
string resp;
using (var response = (HttpWebResponse)webRequest2.EndGetResponse(aResp)) {
using (var streamResponse = response.GetResponseStream()) {
using (var streamReader = new System.IO.StreamReader(streamResponse)) {
resp = streamReader.ReadToEnd();
}
}
}
}, webRequest);
}, req);
One of the issues I couldn't solve though was exceptions thrown when server returns 302 - it seems to throw WebException error with "Not found" description.

// Ensure this is set to true BEFORE navigating to the page
webBrowser1.IsScriptEnabled = true;
// Once the page has loaded, you can read the cookie string
string cookieString = webBrowser1.InvokeScript("eval", new string[] { "document.cookie;" }) as string;
The cookieString variable will contain the full cookie for the document. You can then parse the string.

There is an WebBrowser Extension class which is exactly developed for this:
CookieCollection tempCookies = Microsoft.Phone.Controls.WebBrowserExtensions.GetCookies(this.BrowserControl);