How can I use the FaultContract attribute with RIA Services so that I'll be able to pass additional information as part of an exception to Silverlight?
So I went hunting though the decompiled RIA Services code. It doesn't seem like it's possible to significantly alter the error info that is sent to the client.
You are able to override the OnError() method in your DomainService, but this doesn't allow you to pass arbitrary information back, even if it's a custom exception type.
The reason is buried inside the exception handling of the DomainServices.Hosting.QueryProcessor class.
If an unhandled exception occurs in a domain operation, it bubbles back and then a FaultException() is ultimately thrown (which WCF natively handles).
Unfortuantely, the DomainServiceFault class is very lightweight... It only has a few properties...
public class DomainServiceFault
{
public int ErrorCode { get; set; }
public string ErrorMessage { get; set; }
public bool IsDomainException { get; set; }
public string StackTrace { get; set; }
public IEnumerable<ValidationResultInfo> OperationErrors { get; set; }
public IEnumerable<ValidationResult> GetValidationErrors()
{}
}
and these are populated in the ServiceUtility.CreateFaultExceotion() like so:
DomainServiceFault detail = new DomainServiceFault();
<snip/>
detail.ErrorCode = domainException.ErrorCode;
detail.ErrorMessage = ServiceUtility.FormatExceptionMessage((Exception) domainException);
detail.IsDomainException = true;
if (current != null && !current.IsCustomErrorEnabled)
detail.StackTrace = domainException.StackTrace;
return new FaultException<DomainServiceFault>(detail, new FaultReason(new FaultReasonText(detail.ErrorMessage ?? string.Empty, CultureInfo.CurrentCulture)));
It's worth noting in the case of an Exception, rather than validation errors, the OperationErrors are not populated.
So the upshot of all of this is that I don't believe it's possible to wrap or attach custom exception information to the DomainService error handler (which is really unfortunate).
Related
I'm having a weird issue that I've never seen before, with schema naming using EF core code-first.
I just created a new class LogEntry used to log SMS and emails sent to our users.
public class LogEntry
{
public LogEntry(Guid id)
{
Id = id;
}
public Guid Id { get; set; }
public Provider Provider { get; set; } // Enum
public string Content { get; set; }
public string Recipient { get; set; }
...
}
I then added configuration in my database context class, in OnModelCreating(modelBuilder modelBuilder)
public virtual DbSet<LogEntry> Log { get; set; }
protected override void OnModelCreating(ModelBuilder modelBuilder)
{
modelBuilder.Entity<LogEntry>(entity =>
{
entity.HasKey(log => log.Id);
entity.HasIndex(log => log.Date);
entity.HasIndex(log => log.Provider);
entity.HasIndex(log => log.Recipient);
...
});
}
Then I ran dotnet ef migrations add SomeMigration to actually add the migration. I setup auto migration so it automatically updates my database when the project launches. So far so good.
Now, once I went to check out the new tables it created, I made a weird naming convention regarding the databae schema.
My IIS website application pool is running with a specific managed AD user, let's call it msvc-log-api
I'm used to EF always using the dbo schema, as it default to that schema, but for some weird reason, EF decided to create a new schema named after my managed AD user mydomain\msvc-log-api$. This means that my tables are named in the following way:
mydomain\msvc-log-api$.__EFMigrationHistory
mydomain\msvc-log-api$.Log
Any idea why this is happening, and do I really need to add modelBuilder.HasDefaultSchema("dbo") to mitigate this issue?
The Display.ElementSelected event I get in my Alexa Skill,
how does the event get triggered? And how do Handle it in C#? Can you provide a working example.
I implemented an ElementSelected class in my project and I parse the request i receive from Alexa looking for it, but nothing returns. I don't think I'm doing this correctly. This is my class
namespace Alexa.NET.Request.Type
{
public class ElementSelectedRequest : Request
{
private const string type = "Display.ElementSelected";
protected ElementSelectedRequest()
{
this.Type = type;
}
[JsonProperty("token")]
public string Token { get; set; }
}
}
I've seen all the java and javascript examples on github and I still don't understand how it works exactly.
I tried to solve by myself, but... Looks like I need help from people.
I have Business Silverlight application with WCF RIA and EntityFramework. Access to Database I get via LinqToEntites.
Common loading data from database I making by this:
return DbContext.Customers
This code returns full Customers table from DataBase. But sometimes I do not need to show all data. Easy way is use linq filters in client side by next code:
public LoadInfo()
{
...
var LO1 = PublicDomainContext.Load(PublicDomainContext.GetCustomersQuery());
LO1.Completed += LO1Completed;
...
}
private void LO1Completed(object sender, EventArgs eventArgs)
{
...
DatatViewGrid.ItemsSource = null;
DatatViewGrid.ItemsSource = loadOperation.Entities.Where(c=>c ...filtering...);
//or PublicDomainContext.Customers.Where(c=>c ...filtering...)
...
}
However this way has very and very important flaw: all data passing from server to client side via DomainService may be viewed by applications like Fiddler. So I need to come up with another way.
Task: filter recieving data in server side and return this data.
Way #1: LinqToEntites has a beautiful projection method:
//MSDN Example
var query =
contacts.SelectMany(
contact => orders.Where(order =>
(contact.ContactID == order.Contact.ContactID)
&& order.TotalDue < totalDue)
.Select(order => new
{
ContactID = contact.ContactID,
LastName = contact.LastName,
FirstName = contact.FirstName,
OrderID = order.SalesOrderID,
Total = order.TotalDue
}));
But, unfortunately, DomainServices cannot return undefined types, so this way won't work.
Way #2: I found next solution - make separate DTO classes (DataTransferObject). I just read some samples and made on the server side next class:
[DataContract]
public partial class CustomerDTO
{
[DataMember]
public int ISN { get; set; }
[DataMember]
public string FIO { get; set; }
[DataMember]
public string Listeners { get; set; }
}
And based this class I made a row of methods which return filtered data:
[OperationContract]
public List<CustomerDTO> Customers_Common()
{
return DbContext.Customers....Select(c => new CustomerDTO { ISN = c.ISN, FIO = c.FIO, Listeners = c.Listeners }).ToList();
}
And this works fine, all good...
But, there is strange problem: running application locally does not affect any troubles, but after publishing project on the Web Site, DomainService returns per each method HTTP 500 Error ("Not Found" exception). Of course, I cannot even LogIn into my application. DomainService is dead. If I delete last class and new methods from application and republish - all works fine, but without speacial filtering...
The Question: what I do wrong, why Service is dying with new classes, or tell me another way to solve my trouble. Please.
U P D A T E :
Hey, finally I solved this!
There is an answer: Dynamic query with WCF RIA Services
Your best shot is to find out what is causing the error. For that, override the OnError method on the DomainService like this:
protected override void OnError(DomainServiceErrorInfo errorInfo)
{
/* Log the error info to a file. Don't forget inner exceptions.
*/
base.OnError(errorInfo);
}
This is useful, because only two exceptions will be passed to the client, so if there are a lot of nested inner exceptions, you should still be able to see what actually causes the error.
In addition, you can inspect the error by attaching the debugger to the browser instance you are opening the site with. In VS2010 this is done by doing [Debug] -> [Attach to Process] in the menu-bar.
I have a need to pass an HTTP header for each an every RIA Service request being made from a Silverlight app. The value of the header needs to come from the app instance, and not from a cookie. I know this can be accomplished by putting it in the DTOs, but it's not an option because a lot of our service calls use Entities and change sets, so there's no base class to tie into for all requests. So I'm looking for a centralized and safe means to pass something back with each request so the developers do not have to worry with it. A custom HTTP header would work fine, but I don't know how to intercept the outbound request to set it.
Anyone have any ideas I can try?
On the lower level you can add HTTP headers with the help of an IClientMessageInspector. Try starting from this post on SL forum.
The next step depends on your use cases.
If the value of the header must be the same for any method called by the DomainContext, then you may just extend the context using partial class, add a property for the header value and use that property in the inspector.
If you need to pass a different value for each method call, you'd probably need to wrap your DomainContext into another class and add an argument to each method of the context that will accept the header value and pass it to the inspector somehow. Needless to say, without a code-generator this would be hard.
Here's an adapted sample from the SL forum for the first case:
public sealed partial class MyDomainContext
{
public string HeaderValue { get; set; }
partial void OnCreated()
{
WebDomainClient<IMyDomainServiceContract> webDomainClient = (WebDomainClient<IMyDomainServiceContract>)DomainClient;
CustomHeaderEndpointBehavior customHeaderEndpointBehavior = new CustomHeaderEndpointBehavior(this);
webDomainClient.ChannelFactory.Endpoint.Behaviors.Add(customHeaderEndpointBehavior);
}
}
public class CustomHeaderEndpointBehavior : IEndpointBehavior
{
MyDomainContext _Ctx;
public CustomHeaderEndpointBehavior(MyDomainContext ctx)
{
this._Ctx = ctx;
}
public void AddBindingParameters(ServiceEndpoint endpoint, BindingParameterCollection bindingParameters) { }
public void ApplyClientBehavior(ServiceEndpoint endpoint, System.ServiceModel.Dispatcher.ClientRuntime clientRuntime)
{
clientRuntime.MessageInspectors.Add(new CustomHeaderMessageInspector(this._Ctx));
}
public void ApplyDispatchBehavior(ServiceEndpoint endpoint, System.ServiceModel.Dispatcher.EndpointDispatcher endpointDispatcher) { }
public void Validate(ServiceEndpoint endpoint) { }
}
public class CustomHeaderMessageInspector : IClientMessageInspector
{
MyDomainContext _Ctx;
public CustomHeaderMessageInspector(MyDomainContext ctx)
{
this._Ctx = ctx;
}
public void AfterReceiveReply(ref System.ServiceModel.Channels.Message reply, object correlationState) {}
public object BeforeSendRequest(ref System.ServiceModel.Channels.Message request, IClientChannel channel)
{
string myHeaderName = "X-Foo-Bar";
string myheaderValue = this._Ctx.HeaderValue;
HttpRequestMessageProperty property = (HttpRequestMessageProperty)request.Properties[HttpRequestMessageProperty.Name];
property.Headers[myHeaderName] = myheaderValue;
return null;
}
}
In my Silverlight client I have a partial class created by setting a WCF reference. I've extended this class adding a few RelayCommand properties. I need to initialize these properties which I would normally do in the constructor. However it seems that the constructor is not being called, which I believe is a result of of VTS However I'm also unsuccessful in using the OnDeserialized attribute.
What is the prescribed way to initialize client side data members of a WCF class.
I've created a sample project and everything works as expected. If this code doesn't help - post your data contract and client code.
namespace SilverlightApplication3.ServiceReference1
{
public partial class SomeModel
{
public string ExtendedProperty { get; set; }
[OnDeserializing]
public void OnDeserializingMethod(StreamingContext context)
{
this.ExtendedProperty = "Ok";
}
}
}
Service method call:
var proxy = new ServiceReference1.Service1Client();
proxy.DoWorkCompleted += (s,e) => Debug.WriteLine(e.Result.ExtendedProperty); //Ok
proxy.DoWorkAsync();