how pass parameter from IsAuthorized to HandleUnauthorizedRequest Method web api - c#

I extended AuthorizeAttribute and based on some condition I want display different messages and httpStatusCode to the user. My code is:
protected override bool IsAuthorized(HttpActionContext actionContext)
{
string headerApiKeyValue = string.Empty;
try
{
IEnumerable<string> headerValues;
bool isExitsInHeader = actionContext.Request.Headers.TryGetValues("ApiKey",out headerValues);
if (!isExitsInHeader || headerValues == null)
{
actionContext.Response = CreateResponse(HttpStatusCode.BadRequest, apiKeyNotExistInHeader);
return false;
}
return true;
}
catch (Exception exception)
{
// log exception [apiKeyValue]
return false;
}
}
protected override void HandleUnauthorizedRequest(HttpActionContext actionContext)
{
string a = actionContext.Request.ToString();
// log here
actionContext.Response = CreateResponse(HttpStatusCode.MethodNotAllowed, accessDeniedMessage);
}
In above code there is two response types: First is when the key dos not exist in the header and the second is IsAuthorized method throw exception. How to handle this two types of messages in HandleUnauthorizedRequest method? Is there any way to pass parameter to this method?

update: (I just realized this question is almost 2 years old by now, so I don't assume, we'll here back from the op)
You can just use a private field to store any information you need. Here is a really dead simple example demonstrating this:
public class HeaderAuthenticationAttribute : AuthorizeAttribute
{
private bool invalidApiKey = false;
protected override bool IsAuthorized(HttpActionContext actionContext)
{
bool isExitsInHeader = actionContext.Request.Headers.TryGetValues("ApiKey", out IEnumerable<string> headerValues);
if (isExitsInHeader && headerValues.Any(x => x.Contains("foo")))
{
return true;
}
else if (isExitsInHeader && !headerValues.Any(x => x.Contains("foo")))
{
return false;
}
else
{
invalidApiKey = true;
return false;
}
}
protected override void HandleUnauthorizedRequest(HttpActionContext actionContext)
{
//check if response has header indicating invalid api key
if (invalidApiKey)
{
actionContext.Response = new HttpResponseMessage()
{
StatusCode = HttpStatusCode.MethodNotAllowed,
Content = new StringContent("errorMessage here")
};
}
else
{
actionContext.Response = new HttpResponseMessage()
{
StatusCode = HttpStatusCode.Forbidden,
Content = new StringContent("someOther message here")
};
}
// log here
}
}
What this does (and it is not really a real world example) is the following:
Check if the header "ApiKey" is present
If it is and any of those results is "foo", we're golden
If it is, but no value is "foo", return false
If no header is present, set the invalid header field to true and return false
The HandleUnauthorizedRequest then just checks if this private field is set or not and uses it as a switch to return different HttpResponseMessages with different Status Codes to the user.
In the end you onjly need to implement your logic to validate the api Key and return the appropriate messages as per your requirements.

throw new HttpResponseException(HttpStatusCode.Unauthorized);
Try this & pass the exception you want

Related

C# - Correct approach for method with various validations

I want to know what is the correct way of doing this: lets say I have a login method that receives username and password, and log ins the user or return invalid username/password or not enough permissions. What is the correct way of doing this?
Way #1: throwing exception and handling in the user interface to display the error
public void Login(string username, string password)
{
if (SessionService.GetSession.IsLoggedIn)
{
throw new Exception("User is already logged in");
}
var user = GetByUsername(username);
if (user == null)
{
throw new LoginException(LoginResultEnum.InvalidUsername);
}
var hashPass = EncryptionService.Hash(password);
if (hashPass != user.password)
{
throw new LoginException(LoginResultEnum.InvalidPassword);
}
if (!user.HasPermission(PermissionTypeEnum.CanLogIn))
{
throw new MissingPermissionException(TipoPermisoEnum.CanLogIn);
}
SessionService.GetSession.Login(user);
}
Way #2: returning boolean true/false and handle the error in the UI (success or fail)
public bool Login(string username, string password)
{
if (SessionService.GetSession.IsLoggedIn)
{
return false;
}
var user = GetByUsername(username);
if (user == null)
{
return false;
}
var hashPass = EncryptionService.Hash(password);
if (hashPass != user.password)
{
return false;
}
if (!user.HasPermission(PermissionTypeEnum.CanLogIn))
{
return false;
}
SessionService.GetSession.Login(user);
return true;
}
Way #3: returning a LoginResult enum and handle in the UI
public LoginResult Login(string username, string password)
{
if (SessionService.GetSession.IsLoggedIn)
{
return LoginResult.AlreadyLoggedIn;
}
var user = GetByUsername(username);
if (user == null)
{
return LoginResult.InvalidUsername;
}
var hashPass = EncryptionService.Hash(password);
if (hashPass != user.password)
{
return LoginResult.InvalidPassword;
}
if (!user.HasPermission(PermissionTypeEnum.CanLogIn))
{
return LoginResult.Forbidden;
}
SessionService.GetSession.Login(user);
return LoginResult.OK;
}
In my view it better to create some dto if it is eligible for your case. So this dto will have the following properties:
public class LoginResponseDto
{
public bool Success { get; set; }
public string Error { get; set; }
}
And then you will return your response something like this:
public LoginResponseDto Login(string username, string password)
{
if (SessionService.GetSession.IsLoggedIn)
{
return new LoginResponseDto { Error = "User is already logged in" };
}
var user = GetByUsername(username);
if (user == null)
{
return new LoginResponseDto { Error = "There is no such user" };
}
var hashPass = EncryptionService.Hash(password);
if (hashPass != user.password)
{
return new LoginResponseDto { Error = "Incorrect password or username" };
}
if (!user.HasPermission(PermissionTypeEnum.CanLogIn))
{
return new LoginResponseDto { Error = "There is no permission to log in" };
}
SessionService.GetSession.Login(user);
return new LoginResponseDto { Success = true };
}
It is possible to see this tutorial "Create a secure ASP.NET MVC 5 web app with log in, email confirmation and password reset". Author of article use ViewBag in this article to send errors from controller and Succeeded to check whether login is okay.
In addition, try to avoid to show message about what is exactly wrong username or password.
I would say #3 is the best way.
#1 you are using Exception for non-exceptional circumstances. The control path is expected, so don't use Exceptions.
#2 By using a bool you are discarding information, is it InvalidPassword or Forbidden?
#3 Returns all information, allowing the UI to surface that information to the User.

Implied property of type Object in C# - is it possible?

I'm trying to convert the following function from VB to C#:
Private Function CheckHeader(ByVal Request As Object, ByRef Response As Object) As Boolean
CheckHeader = True
If Request.Header Is Nothing Then
CheckHeader = False
Throw New System.ArgumentException("Header Object not found")
End If
End Function
And here is what I have so far in C#:
private bool CheckHeader(object Request, ref object Response)
{
bool functionReturnValue = false;
functionReturnValue = true;
var localRequest = Request;
if (localRequest.Header == null)
{
functionReturnValue = false;
throw new System.ArgumentException("Header Object not found");
}
return functionReturnValue;
}
The problem is with the Request parameter that is of type Object. In VB, we have the nice feature of the implied property, which of course C# doesn't like one bit. Any object that does get passed as the "Request" parameter is supposed to have the Header parameter. I'm just checking to see if it's there. I'm thinking the use of generics might apply here, but I'm unsure of the route to take. Any ideas?
only way I can think for you to do the same thing in c# is to use the dynamic keyword
it should behave like the strict off of vb.net
private bool CheckHeader(dynamic Request, ref object Response)
{
bool functionReturnValue = false;
functionReturnValue = true; //you should put true above instead
var localRequest = Request; //not sure why this but whatever
if (localRequest.Header == null)
{
functionReturnValue = false; //no need for this here, because of the throw
throw new System.ArgumentException("Header Object not found");
}
return functionReturnValue; //at this point remove
//this variable and simply return true
}
so this should be ok
private bool CheckHeader(dynamic Request, ref object Response)
{
//Response variable should stay there since it might be a
//breaking change, you should check if you can remove it
if (Request.Header == null)
{
throw new System.ArgumentException("Header Object not found");
}
return true;
}
Because object has no Header property, if possible I would convert Request Object into a c# interface. So I would change the code to look like this:
First IRequest interface:
public interface IRequest
{
//Assuming Header is an object.
MyHeaderObject Header {get;set;}
}
Now CheckHeader function:
private bool CheckHeader(IRequest Request, ref object Response)
{
bool functionReturnValue = false;
functionReturnValue = true;
var localRequest = Request;
if (localRequest.Header == null)
{
functionReturnValue = false;
throw new System.ArgumentException("Header Object not found");
}
return functionReturnValue;
}
Now everything that passes a Request to CheckHeader needs to implement IRequest.
Here is an example of it:
public MyRequest : IRequest
{
public MyHeaderObject Header {get;set;}
}

Return error from models

I have the following model:
internal static List<Contracts.DataContracts.Report> GetReportsForSearch(string searchVal, string searchParam)
{
var param1 = new SqlParameter("#SearchVal", searchVal);
var ctx = new StradaDataReviewContext2();
var reports = new List<Contracts.DataContracts.Report>();
try
{
//Validate param1 here and return false if the requirment are not met
}
catch(Exception e)
{
//Throw
}
}
param1 here Is a value entered by a user and I want to validate It here, and If the requirements are not met, I want to return an error.
But how can I return an error here from the model? The method Is of the type List, and I can't not just write return false in this method.
Any suggestion how to do It?
It is good that you didn't thought about throwing an exception, when requirements are not met. We shouldn't use exceptions for controlling program flow.
I have two options in my mind :
1. Use objects
Modify your GetReportsForSearch method to following signature:
internal static List<Contracts.DataContracts.Report> GetReportsForSearch(string searchVal,
string searchParam, ReportRequestor requestor)
{
var param1 = new SqlParameter("#SearchVal", searchVal);
var ctx = new StradaDataReviewContext2();
var reports = new List<Contracts.DataContracts.Report>();
try
{
//Validate param1 here and call RequirementsAreNotMet method if the requirements are not met
requestor.RequirementsAreNotMet();
}
catch(Exception e)
{
//Throw
}
}
And then you can implement code responsible for handling this situation in ReportRequestor class
public class ReportRequestor
{
public void RequiremenrsAreNotMet()
{
//code which handle situation when requiremenets are not met
}
}
2. Use return type as indicator of status
In this way, when requirements are not met you should create ReportGenerationStatus object with HasResult flag set to false.
In other case just set HasResult to true and also set results accordingly. This somewhat mimics Option type known from functional languages
internal static ReportGenerationStatus GetReportsForSearch(string searchVal, string searchParam)
{
//code for your method
}
public class ReportGenerationStatus
{
public List<Contracts.DataContracts.Report> Result { get; set; }
public bool HasResult { get; set; }
}

WebAPI Controller Ignoring CORS

I have a WebAPI controller with a custom CORS policy provider attribute on the class. In defining the attribute, I have the following constructor.
[AttributeUsage(AttributeTargets.Method | AttributeTargets.Class, AllowMultiple = false)]
public class ConfiguredCORSPolicyProviderAttribute : ActionFilterAttribute, ICorsPolicyProvider
{
private CorsPolicy _policy;
public ConfiguredCORSPolicyProviderAttribute()
{
_policy = new CorsPolicy
{
AllowAnyMethod = true,
AllowAnyHeader = true
};
// If there are no domains in the 'CORSDomainSection' section in Web.config, all origins will be allowed by default.
var domains = (CORSDomainSection)ConfigurationManager.GetSection("CORSDomainSection");
if (domains != null)
{
foreach (DomainConfigElement domain in domains.Domains)
{
_policy.Origins.Add(domain.Domain);
}
}
else
{
_policy.AllowAnyOrigin = true;
}
}
public Task<CorsPolicy> GetCorsPolicyAsync(HttpRequestMessage request)
{
return Task.FromResult(_policy);
}
public Task<CorsPolicy> GetCorsPolicyAsync(HttpRequestMessage request, CancellationToken token)
{
return GetCorsPolicyAsync(request);
}
}
The ConfigurationManager gets a list (from Web.config) of acceptable origins/domains that I want to allow to make requests.
This code appropriately handles the "Access-Control-Allow-Origin" header, adding it when the request origin is on the list, and withholding it when not. However, the code in the controller still gets called no matter what.
Why, and how do I appropriately prevent the controller from executing if the origin of the request isn't allowed?
UPDATE || [April 12, 2016 # 12:30p]
I was able to resolve the issue using a combination of OnActionExecuted and OnActionExecuting method overrides, code below.
/// <summary>
/// Executed after the action method is invoked.
/// </summary>
/// <param name="context">The context of the HTTP request.</param>
public override void OnActionExecuted(HttpActionExecutedContext context)
{
string requestOrigin;
try
{
requestOrigin = context.Request.Headers.GetValues("Origin").FirstOrDefault();
}
catch
{
requestOrigin = string.Empty;
}
if (IsAllowedOrigin(requestOrigin))
{
context.Response.Headers.Add("Access-Control-Allow-Origin", requestOrigin);
if (IsPreflight(context))
{
string allowedMethods = string.Empty;
string allowedHeaders = string.Empty;
if (Policy.AllowAnyMethod)
{
allowedMethods = context.Request.Headers.GetValues("Access-Control-Request-Method").FirstOrDefault();
}
else
{
foreach (var method in Policy.Methods)
{
if (Policy.Methods.IndexOf(method) == 0)
{
allowedMethods = method;
}
else
{
allowedMethods += string.Format(", {0}", method);
}
}
}
try
{
if (Policy.AllowAnyHeader)
{
allowedHeaders = context.Request.Headers.GetValues("Access-Control-Request-Headers").FirstOrDefault();
}
else
{
foreach (var header in Policy.Headers)
{
if (Policy.Headers.IndexOf(header) == 0)
{
allowedHeaders = header;
}
else
{
allowedHeaders += string.Format(", {0}", header);
}
}
}
context.Response.Headers.Add("Access-Control-Allow-Headers", allowedHeaders);
}
catch
{
// Do nothing.
}
context.Response.Headers.Add("Access-Control-Allow-Methods", allowedMethods);
}
}
base.OnActionExecuted(context);
}
/// <summary>
/// Executed before the action method is invoked.
/// </summary>
/// <param name="context">The context of the HTTP request.</param>
public override void OnActionExecuting(HttpActionContext context)
{
string requestOrigin;
try
{
requestOrigin = context.Request.Headers.GetValues("Origin").FirstOrDefault();
}
catch
{
requestOrigin = string.Empty;
}
if (IsAllowedOrigin(requestOrigin))
{
base.OnActionExecuting(context);
}
else
{
context.ModelState.AddModelError("State", "The origin of the request is forbidden from making requests.");
context.Response = context.Request.CreateErrorResponse(HttpStatusCode.Forbidden, context.ModelState);
}
}
private bool IsAllowedOrigin(string requestOrigin)
{
requestOrigin = requestOrigin.Replace("https://", "").Replace("http://", "");
if (System.Diagnostics.Debugger.IsAttached || PolicyContains(requestOrigin))
{
return true;
}
else
{
return false;
}
}
private bool PolicyContains(string requestOrigin)
{
foreach (var domain in _policy.Origins)
{
if (domain.Replace("https://", "").Replace("http://", "") == requestOrigin)
{
return true;
}
}
return false;
}
private bool IsPreflight(HttpActionExecutedContext context)
{
string header = string.Empty;
try
{
header = context.Request.Headers.GetValues("Access-Control-Request-Method").FirstOrDefault();
}
catch
{
return false;
}
if (header != null && context.Request.Method == HttpMethod.Options)
{
return true;
}
else
{
return false;
}
}
CORS headers are not expected to prevent calls to controller - i.e. if native client (pretty much anything that is not a browser) calls the method it should be handled by default.
If you really need to block such calls - perform similar checks in before controller get called in OnActionExecutingAsync.

How to use ETag in Web API using action filter along with HttpResponseMessage

I have a ASP.Net Web API controller which simply returns the list of users.
public sealed class UserController : ApiController
{
[EnableTag]
public HttpResponseMessage Get()
{
var userList= this.RetrieveUserList(); // This will return list of users
this.responseMessage = new HttpResponseMessage(HttpStatusCode.OK)
{
Content = new ObjectContent<List<UserViewModel>>(userList, new JsonMediaTypeFormatter())
};
return this.responseMessage;
}
}
and an action filter attribute class EnableTag which is responsible to manage ETag and cache:
public class EnableTag : System.Web.Http.Filters.ActionFilterAttribute
{
private static ConcurrentDictionary<string, EntityTagHeaderValue> etags = new ConcurrentDictionary<string, EntityTagHeaderValue>();
public override void OnActionExecuting(HttpActionContext context)
{
if (context != null)
{
var request = context.Request;
if (request.Method == HttpMethod.Get)
{
var key = GetKey(request);
ICollection<EntityTagHeaderValue> etagsFromClient = request.Headers.IfNoneMatch;
if (etagsFromClient.Count > 0)
{
EntityTagHeaderValue etag = null;
if (etags.TryGetValue(key, out etag) && etagsFromClient.Any(t => t.Tag == etag.Tag))
{
context.Response = new HttpResponseMessage(HttpStatusCode.NotModified);
SetCacheControl(context.Response);
}
}
}
}
}
public override void OnActionExecuted(HttpActionExecutedContext context)
{
var request = context.Request;
var key = GetKey(request);
EntityTagHeaderValue etag;
if (!etags.TryGetValue(key, out etag) || request.Method == HttpMethod.Put || request.Method == HttpMethod.Post)
{
etag = new EntityTagHeaderValue("\"" + Guid.NewGuid().ToString() + "\"");
etags.AddOrUpdate(key, etag, (k, val) => etag);
}
context.Response.Headers.ETag = etag;
SetCacheControl(context.Response);
}
private static void SetCacheControl(HttpResponseMessage response)
{
response.Headers.CacheControl = new CacheControlHeaderValue()
{
MaxAge = TimeSpan.FromSeconds(60),
MustRevalidate = true,
Private = true
};
}
private static string GetKey(HttpRequestMessage request)
{
return request.RequestUri.ToString();
}
}
The above code create an attribute class to manage ETag. So on the first request, it will create a new E-Tag and for the subsequent request it will check whether any ETag is existed. If so, it will generate Not Modified HTTP Status and return back to client.
My problem is, I want to create a new ETag if there are changes in my user list, ex. a new user is added, or an existing user is deleted. and append it with the response. This can be tracked by the userList variable.
Currently, the ETag received from client and server are same from every second request, so in this case it will always generate Not Modified status, while I want it when actually nothing changed.
Can anyone guide me in this direction?
My requirement was to cache my web api JSON responses... And all the solutions provided don't have an easy "link" to where the data is generated - ie in the Controller...
So my solution was to create a wrapper "CacheableJsonResult" which generated a Response, and then added the ETag to the header. This allows a etag to be passed in when the controller method is generated and wants to return the content...
public class CacheableJsonResult<T> : JsonResult<T>
{
private readonly string _eTag;
private const int MaxAge = 10; //10 seconds between requests so it doesn't even check the eTag!
public CacheableJsonResult(T content, JsonSerializerSettings serializerSettings, Encoding encoding, HttpRequestMessage request, string eTag)
:base(content, serializerSettings, encoding, request)
{
_eTag = eTag;
}
public override Task<HttpResponseMessage> ExecuteAsync(System.Threading.CancellationToken cancellationToken)
{
Task<HttpResponseMessage> response = base.ExecuteAsync(cancellationToken);
return response.ContinueWith<HttpResponseMessage>((prior) =>
{
HttpResponseMessage message = prior.Result;
message.Headers.ETag = new EntityTagHeaderValue(String.Format("\"{0}\"", _eTag));
message.Headers.CacheControl = new CacheControlHeaderValue
{
Public = true,
MaxAge = TimeSpan.FromSeconds(MaxAge)
};
return message;
}, cancellationToken);
}
}
And then, in your controller - return this object:
[HttpGet]
[Route("results/{runId}")]
public async Task<IHttpActionResult> GetRunResults(int runId)
{
//Is the current cache key in our cache?
//Yes - return 304
//No - get data - and update CacheKeys
string tag = GetETag(Request);
string cacheTag = GetCacheTag("GetRunResults"); //you need to implement this map - or use Redis if multiple web servers
if (tag == cacheTag )
return new StatusCodeResult(HttpStatusCode.NotModified, Request);
//Build data, and update Cache...
string newTag = "123"; //however you define this - I have a DB auto-inc ID on my messages
//Call our new CacheableJsonResult - and assign the new cache tag
return new CacheableJsonResult<WebsiteRunResults>(results, GlobalConfiguration.Configuration.Formatters.JsonFormatter.SerializerSettings, System.Text.UTF8Encoding.Default, Request, newTag);
}
}
private static string GetETag(HttpRequestMessage request)
{
IEnumerable<string> values = null;
if (request.Headers.TryGetValues("If-None-Match", out values))
return new EntityTagHeaderValue(values.FirstOrDefault()).Tag;
return null;
}
You need to define how granular to make your tags; my data is user-specific, so I include the UserId in the CacheKey (etag)
a good solution for ETag and in ASP.NET Web API is to use CacheCow . A good article is here.
It's easy to use and you don't have to create a custom Attribute.
Have fun
.u
I found CacheCow very bloated for what it does, if the only reason is, to lower the amount of data transfered, you might want to use something like this:
public class EntityTagContentHashAttribute : ActionFilterAttribute
{
private IEnumerable<string> _receivedEntityTags;
private readonly HttpMethod[] _supportedRequestMethods = {
HttpMethod.Get,
HttpMethod.Head
};
public override void OnActionExecuting(HttpActionContext context) {
if (!_supportedRequestMethods.Contains(context.Request.Method))
throw new HttpResponseException(context.Request.CreateErrorResponse(HttpStatusCode.PreconditionFailed,
"This request method is not supported in combination with ETag."));
var conditions = context.Request.Headers.IfNoneMatch;
if (conditions != null) {
_receivedEntityTags = conditions.Select(t => t.Tag.Trim('"'));
}
}
public override void OnActionExecuted(HttpActionExecutedContext context)
{
var objectContent = context.Response.Content as ObjectContent;
if (objectContent == null) return;
var computedEntityTag = ComputeHash(objectContent.Value);
if (_receivedEntityTags.Contains(computedEntityTag))
{
context.Response.StatusCode = HttpStatusCode.NotModified;
context.Response.Content = null;
}
context.Response.Headers.ETag = new EntityTagHeaderValue("\"" + computedEntityTag + "\"", true);
}
private static string ComputeHash(object instance) {
var cryptoServiceProvider = new MD5CryptoServiceProvider();
var serializer = new DataContractSerializer(instance.GetType());
using (var memoryStream = new MemoryStream())
{
serializer.WriteObject(memoryStream, instance);
cryptoServiceProvider.ComputeHash(memoryStream.ToArray());
return String.Join("", cryptoServiceProvider.Hash.Select(c => c.ToString("x2")));
}
}
}
No need for setting up anything, set and forget. The way i like it. :)
I like the answer which was provided by #Viezevingertjes. It is the most elegant and "No need for setting up anything" approach is very convenient. I like it too :)
However I think it has a few drawbacks:
The whole OnActionExecuting() method and storing ETags in _receivedEntityTags is unnecessary because the Request is available inside the OnActionExecuted method as well.
Only works with ObjectContent response types.
Extra work load because of the serialization.
Also it was not part of the question and nobody mentioned it. But ETag should be used for Cache validation. Therefore it should be used with Cache-Control header so clients don't even have to call the server until the cache expires (it can be very short period of time depends on your resource). When the cache expired then client makes a request with ETag and validate it. For more details about caching see this article.
So that's why I decided to pimp it up a little but. Simplified filter no need for OnActionExecuting method, works with Any response types, no Serialization. And most importantly adds CacheControl header as well. It can be improved e.g. with Public cache enabled, etc... However I strongly advise you to understand caching and modify it carefully. If you use HTTPS and the endpoints are secured then this setup should be fine.
/// <summary>
/// Enables HTTP Response CacheControl management with ETag values.
/// </summary>
public class ClientCacheWithEtagAttribute : ActionFilterAttribute
{
private readonly TimeSpan _clientCache;
private readonly HttpMethod[] _supportedRequestMethods = {
HttpMethod.Get,
HttpMethod.Head
};
/// <summary>
/// Default constructor
/// </summary>
/// <param name="clientCacheInSeconds">Indicates for how long the client should cache the response. The value is in seconds</param>
public ClientCacheWithEtagAttribute(int clientCacheInSeconds)
{
_clientCache = TimeSpan.FromSeconds(clientCacheInSeconds);
}
public override async Task OnActionExecutedAsync(HttpActionExecutedContext actionExecutedContext, CancellationToken cancellationToken)
{
if (!_supportedRequestMethods.Contains(actionExecutedContext.Request.Method))
{
return;
}
if (actionExecutedContext.Response?.Content == null)
{
return;
}
var body = await actionExecutedContext.Response.Content.ReadAsStringAsync();
if (body == null)
{
return;
}
var computedEntityTag = GetETag(Encoding.UTF8.GetBytes(body));
if (actionExecutedContext.Request.Headers.IfNoneMatch.Any()
&& actionExecutedContext.Request.Headers.IfNoneMatch.First().Tag.Trim('"').Equals(computedEntityTag, StringComparison.InvariantCultureIgnoreCase))
{
actionExecutedContext.Response.StatusCode = HttpStatusCode.NotModified;
actionExecutedContext.Response.Content = null;
}
var cacheControlHeader = new CacheControlHeaderValue
{
Private = true,
MaxAge = _clientCache
};
actionExecutedContext.Response.Headers.ETag = new EntityTagHeaderValue($"\"{computedEntityTag}\"", false);
actionExecutedContext.Response.Headers.CacheControl = cacheControlHeader;
}
private static string GetETag(byte[] contentBytes)
{
using (var md5 = MD5.Create())
{
var hash = md5.ComputeHash(contentBytes);
string hex = BitConverter.ToString(hash);
return hex.Replace("-", "");
}
}
}
Usage e.g: with 1 min client side caching:
[ClientCacheWithEtag(60)]
Seems to be a nice way to do it:
public class CacheControlAttribute : System.Web.Http.Filters.ActionFilterAttribute
{
public int MaxAge { get; set; }
public CacheControlAttribute()
{
MaxAge = 3600;
}
public override void OnActionExecuted(HttpActionExecutedContext context)
{
if (context.Response != null)
{
context.Response.Headers.CacheControl = new CacheControlHeaderValue
{
Public = true,
MaxAge = TimeSpan.FromSeconds(MaxAge)
};
context.Response.Headers.ETag = new EntityTagHeaderValue(string.Concat("\"", context.Response.Content.ReadAsStringAsync().Result.GetHashCode(), "\""),true);
}
base.OnActionExecuted(context);
}
}

Categories

Resources