We're using ASP.NET WebAPI with Entity Framework (with lazy loading) and using Json.NET for serializing the data to JSON before returning the data to the client.
We are experiencing intermittent sudden spikes in memory usage which we suspect might originate with Json.NET not recognizing reference loops when serializing data (since Entity Framework might be doing some lazy loading voodoo with proxy classes which goes under the radar of Json.NET).
I thought I'd limit how deep Json.NET was allowed to go to serialize data (at least then we'd get a sensible exception when this happens so we could fix it in the data model), but I soon discovered that the MaxDepth property of JsonSerializerSettings only kicks in when DEserializing objects.
Is there any known way of imposing a limit on Json.NET when serializing?
I can't think of a way to do this out-of-the-box with Json.NET, since (as you correctly observe) MaxDepth is ignored when serializing. What you could do is to subclass JsonTextWriter and do the checks yourself:
public class MaxDepthJsonTextWriter : JsonTextWriter
{
public int? MaxDepth { get; set; }
public int MaxObservedDepth { get; private set; }
public MaxDepthJsonTextWriter(TextWriter writer, JsonSerializerSettings settings)
: base(writer)
{
this.MaxDepth = (settings == null ? null : settings.MaxDepth);
this.MaxObservedDepth = 0;
}
public MaxDepthJsonTextWriter(TextWriter writer, int? maxDepth)
: base(writer)
{
this.MaxDepth = maxDepth;
}
public override void WriteStartArray()
{
base.WriteStartArray();
CheckDepth();
}
public override void WriteStartConstructor(string name)
{
base.WriteStartConstructor(name);
CheckDepth();
}
public override void WriteStartObject()
{
base.WriteStartObject();
CheckDepth();
}
private void CheckDepth()
{
MaxObservedDepth = Math.Max(MaxObservedDepth, Top);
if (Top > MaxDepth)
throw new JsonSerializationException(string.Format("Depth {0} Exceeds MaxDepth {1} at path \"{2}\"", Top, MaxDepth, Path));
}
}
Then, to manually generate a JSON string, you would use it like this:
var settings = new JsonSerializerSettings { MaxDepth = 10 };
string json;
try
{
using (var writer = new StringWriter())
{
using (var jsonWriter = new MaxDepthJsonTextWriter(writer, settings))
{
JsonSerializer.Create(settings).Serialize(jsonWriter, myClass);
// Log the MaxObservedDepth here, if you want to.
}
json = writer.ToString();
}
Debug.WriteLine(json);
}
catch (Exception ex)
{
Debug.WriteLine(ex);
throw;
}
Demo fiddle here.
Since your tags include web-api, if you want to do this check inside web API calls, you could follow Rick Strahl's instructions to create a custom MediaTypeFormatter for JSON: Using an alternate JSON Serializer in ASP.NET Web API; then use the code above in the OnWriteToStreamAsync method when generating the json string.
Related
We have an ASP.NET application. We cannot edit source code of controllers. But we can implement ActionFilter.
One of our controller action methods returns JSON. Is it possible to modify it in ActionFilter? We need to add one more property to a returned object.
Maybe, some other way to achieve it?
Found this interesting and as #Chris mentioned, though conceptually I knew this would work, I never tried this and hence thought of giving it a shot. I'm not sure whether this is an elegant/correct way of doing it, but this worked for me. (I'm trying to add Age property dynamically using ActionResult)
[PropertyInjector("Age", 12)]
public ActionResult Index()
{
return Json(new { Name = "Hello World" }, JsonRequestBehavior.AllowGet);
}
And the filter:
public class PropertyInjector : ActionFilterAttribute
{
string key;
object value;
public PropertyInjector(string key, object value)
{
this.key = key;
this.value = value;
}
public override void OnActionExecuted(ActionExecutedContext filterContext)
{
var jsonData = ((JsonResult)filterContext.Result).Data;
JObject data = JObject.FromObject(jsonData);
data.Add(this.key,JToken.FromObject(this.value));
filterContext.Result = new ContentResult { Content = data.ToString(), ContentType = "application/json" };
base.OnActionExecuted(filterContext);
}
}
Update
If it's not dynamic data which is to be injected, then remove filter constructor and hard code key & value directly and then the filter could be registered globally without editing the controller
GlobalFilters.Filters.Add(new PropertyInjector());
Using AutoFixture with the AutoFixture.AutoMoq package, I sometimes find tests that weren't configured to correctly test the thing they meant to test, but the problem was never discovered because of the default (Loose) Mock behavior:
public interface IService
{
bool IsSomethingTrue(int id);
}
void Main()
{
var fixture = new Fixture()
.Customize(new AutoMoqCustomization());
var service = fixture.Freeze<Mock<IService>>();
Console.WriteLine(service.Object.IsSomethingTrue(1)); // false
}
I'd like to make Mocks get created with Strict behavior, so we're forced to call Setup() for the methods we expect to be called. I can do this for each individual mock like this:
fixture.Customize<Mock<IService>>(c => c.FromFactory(() => new Mock<IService>(MockBehavior.Strict)));
But after combing through source code for AutoMoqCustomization() and the various ISpecimenBuilder and other implementations, I'm pretty lost as to the best way to just make all Mocks get initialized with strict behavior. The framework appears to be very flexible and extensible, so I'm sure there's a simple way to do this--I just can't figure out how.
There's no simple built-in feature that will enable you to do something like that, but it shouldn't be that hard to do.
Essentially, you'd need to change MockConstructorQuery so that it invokes the constructor that takes a MockBehavior value, and pass in MockBehavior.Strict.
Now, you can't change that behaviour in MockConstructorQuery, but that class is only some 9-10 lines of code, so you should be able to create a new class that implements IMethodQuery by using MockConstructorQuery as a starting point.
Likewise, you'll also need to create a custom ICustomization that does almost exactly the same as AutoMoqCustomization, with the only exception that it uses your custom IMethodQuery with strict mock configuration instead of MockConstructorQuery. That's another 7 lines of code you'll need to write.
All that said, in my experience, using strict mocks is a bad idea. It'll make your tests brittle, and you'll waste a lot of time mending 'broken' tests. I can only recommend that you don't do this, but now I've warned you; it's your foot.
For those interested, down below you can find #MarkSeemann's reply translated into code. I am pretty sure it does not cover all use cases and it was not heavily tested. But it should be a good starting point.
using System;
using System.Collections.Generic;
using System.Linq;
using System.Reflection;
using Moq;
using Ploeh.AutoFixture;
using Ploeh.AutoFixture.AutoMoq;
using Ploeh.AutoFixture.Kernel;
namespace ConsoleApplication1
{
public class StrictAutoMoqCustomization : ICustomization
{
public StrictAutoMoqCustomization() : this(new MockRelay()) { }
public StrictAutoMoqCustomization(ISpecimenBuilder relay)
{
// TODO Null check params
Relay = relay;
}
public ISpecimenBuilder Relay { get; }
public void Customize(IFixture fixture)
{
// TODO Null check params
fixture.Customizations.Add(new MockPostprocessor(new MethodInvoker(new StrictMockConstructorQuery())));
fixture.ResidueCollectors.Add(Relay);
}
}
public class StrictMockConstructorMethod : IMethod
{
private readonly ConstructorInfo ctor;
private readonly ParameterInfo[] paramInfos;
public StrictMockConstructorMethod(ConstructorInfo ctor, ParameterInfo[] paramInfos)
{
// TODO Null check params
this.ctor = ctor;
this.paramInfos = paramInfos;
}
public IEnumerable<ParameterInfo> Parameters => paramInfos;
public object Invoke(IEnumerable<object> parameters) => ctor.Invoke(parameters?.ToArray() ?? new object[] { });
}
public class StrictMockConstructorQuery : IMethodQuery
{
public IEnumerable<IMethod> SelectMethods(Type type)
{
if (!IsMock(type))
{
return Enumerable.Empty<IMethod>();
}
if (!GetMockedType(type).IsInterface && !IsDelegate(type))
{
return Enumerable.Empty<IMethod>();
}
var ctor = type.GetConstructor(new[] { typeof(MockBehavior) });
return new IMethod[]
{
new StrictMockConstructorMethod(ctor, ctor.GetParameters())
};
}
private static bool IsMock(Type type)
{
return type != null && type.IsGenericType && typeof(Mock<>).IsAssignableFrom(type.GetGenericTypeDefinition()) && !GetMockedType(type).IsGenericParameter;
}
private static Type GetMockedType(Type type)
{
return type.GetGenericArguments().Single();
}
internal static bool IsDelegate(Type type)
{
return typeof(MulticastDelegate).IsAssignableFrom(type.BaseType);
}
}
}
Usage
var fixture = new Fixture().Customize(new StrictAutoMoqCustomization());
I am looking at implementing a logging mechanism in a site of mine, I wish to do basic user action logging. I don't want to log every single button they click on, but I do want to log actions they do which makes changes.
Are there any libraries or articles / tutorials out there which can help me implement a good and efficient logging mechanism for my asp.net site. Im not sure if there are any changes in MVC5 that might come in use for logging as I know user Authentication and Authorization has changed a fair amount from 4 to 5.
I'm sure that there is a dynamic library out there that will work in many different situations.
Nice to haves:
Async capability
Scalable
Simple to use
I'm thinking along the lines of creating a custom filter or attribute that then logs the suers action, but that's just my Idea, Im here to ask what the standard / industry way to do it is.
There isn't an industry standard.
I've used filters or I've overridden the "onActionExecuting" method on the base controller class to record controller / action events.
EDIT ::
Trying to be more helpful but this is really vague.
If you're worried about errors and things like that use elmah.
For other logging use Nlog or Log4Net.
If you're trying to do extra logging like auditing or something like that you can use any combination of those, or something custom. In my site I created a table that stores every click on the site by creating an object sort of like this :
public class audit
{
public int ID { get; set; }
public DateTime AuditDate { get; set; }
public string ControllerName { get; set; }
public string ActionName { get; set; }
public Dictionary<string, object> values
}
In my base constructor, I overrode the OnActionExecuting event :
protected override void OnActionExecuting(ActionExecutingContext ctx)
{
checkForLogging(ctx);
//do not omit this line
base.OnActionExecuting(ctx);
}
Lets say I want to log all Get Requests using my new audit object
private void checkForLogging(ActionExecutingContext ctx)
{
//we leave logging of posts up to the actual methods because they're more complex...
if (ctx.HttpContext.Request.RequestType == "GET")
{
logging(ctx.ActionDescriptor.ActionName, ctx.ActionDescriptor.ControllerDescriptor.ControllerName, ctx.ActionParameters);
}
}
Thats all the info I need to fill my logging object with the action name, the controller name and all the params passed into the method. You can either save this to a db, or a logfile or whatever you want really.
The point is just its a pretty big thing. This is just one way to do it and it may or may not help you. Maybe define a bit more what exactly you want to log and when you want to do it?
You can create a custom attribute and decorate methods with it and then check if that attribute is present when the OnActionExecuting method fires. You can then get that filter if present and read from it and use that to drive your logging if you want...
Maybe this example will help.
My focus on logging is in the CREATE, EDIT, DELETE actions.
I am using MVC 5 Code-first EF 6.1 (VS 2013) ,
and for this example I are referring to the Create action for an entity called "WorkFlow"
I actually view these logs from SSRS, but you could add a controller and Views for WriteUsageLog and view them from the MVC application
MODEL: Create a MODEL Entity called "WriteUsageLog" which will be where the log records are kept
CONTROLLER: Extract, or refactor, the HttpPost overload of the "Create" action from the WorkFlowController into a Partial Class called "WorkFlowController" (These partials are to avoid being deleted and rebuilt when I use the wizard to create Controllers)
Other Classes in the CONTROLLER folder: Then there are some helper functions that are required in a class called "General_Object_Extensions" and "General_ActiveDirectory_Extensions" (NOTE: these are not really 'extensions')
Add the following line to the DBContext:
public DbSet WriteUsageLogs { get; set; }
The advantage of this example is:
I am recording the following for the record:
User Name from Active Directory
The DateTime that the log record is being created
The computer name
And a Note that consists of the values for all the entity properties
I am recording the log in a table from which I can access it either using an MVC controller, or preferably from SQL Server Report Server. Where I can monitor all my MVC applications
/Models/WriteUsageLog.cs
using System;
using System.ComponentModel.DataAnnotations;
using System.ComponentModel.DataAnnotations.Schema;
namespace MileageReimbursement.Models
{
public class WriteUsageLog
{
public WriteUsageLog()
{
this.DateTimeCreated = DateTime.Now; // auto-populates DateTimeCreated field
}
[Key]
public int WriteUsageLogID { get; set; }
[Column(TypeName = "nvarchar(max)")]
public string Note { get; set; }
public string UserLogIn { get; set; }
public string ComputerName { get; set; }
public DateTime DateTimeCreated { get; private set; } //private set to for auto-populates DateTimeCreated field
}
}
/Controllers/ControllerPartials.cs
using System.Linq;
using System.Web.Mvc;
using MileageReimbursement.Models;
//These partials are to avoid be deleted and rebuilt when I use the wizard to create Controllers
namespace MileageReimbursement.Controllers
{
public partial class WorkFlowController : Controller
{
[HttpPost]
[ValidateAntiForgeryToken]
public ActionResult Create([Bind(Include = "whatever")] WorkFlow workFlow)
{
...
if (ModelState.IsValid)
{
db.WorkFlows.Add(workFlow);
db.SaveChanges();
//===================================================================================================================
string sX = workFlow.GetStringWith_RecordProperties();
//===================================================================================================================
var logRecord = new WriteUsageLog();
logRecord.Note = "New WorkFlow Record Added - " + sX;
logRecord.UserLogIn = General_ActiveDirectory_Extensions.fn_sUser();
string IP = Request.UserHostName;
logRecord.ComputerName = General_functions.fn_ComputerName(IP);
db.WriteUsageLogs.Add(logRecord);
db.SaveChanges();
//===================================================================================================================
return RedirectToAction("Index");
}
else // OR the user is directed back to the validation error messages and given an opportunity to correct them
{
...
return View(workFlow); //This sends the user back to the CREATE view to deal with their errors
}
}
}
}
/Controllers/ControllerExtensions.cs
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.DirectoryServices.AccountManagement;
using System.Linq;
using System.Net;
using System.Net.Mail;
using System.Reflection;
using System.Security.Cryptography;
using System.Text;
using System.Web;
namespace MileageReimbursement.Controllers
{
public static class General_ActiveDirectory_Extensions
{
public static string fn_sUser()
{
char cX = '\\';
string sUser = General_functions.fn_ReturnPortionOfStringAfterLastOccuranceOfACharacter(HttpContext.Current.User.Identity.Name, cX);
return sUser; //returns just the short logon name Example for 'accessiicarewnc\ggarson', it returns 'ggarson'
}
} //General_ActiveDirectory_Extensions
public static class General_Object_Extensions
{
public static string GetStringWith_RecordProperties(this object Record)
{
string sX = null;
Dictionary<string, object> _record = GetDictionary_WithPropertiesForOneRecord(Record);
int iPropertyCounter = 0;
foreach (var KeyValuePair in _record)
{
iPropertyCounter += 1;
object thePropertyValue = _record[KeyValuePair.Key];
if (thePropertyValue != null)
{
sX = sX + iPropertyCounter + ") Property: {" + KeyValuePair.Key + "} = [" + thePropertyValue + "] \r\n";
}
else
{
sX = sX + iPropertyCounter + ") Property: {" + KeyValuePair.Key + "} = [{NULL}] \r\n";
}
}
return sX;
}
public static Dictionary<string, object> GetDictionary_WithPropertiesForOneRecord(object atype)
{
if (atype == null) return new Dictionary<string, object>();
Type t = atype.GetType();
PropertyInfo[] props = t.GetProperties();
Dictionary<string, object> dict = new Dictionary<string, object>();
foreach (PropertyInfo prp in props)
{
object value = prp.GetValue(atype, new object[] { });
dict.Add(prp.Name, value);
}
return dict;
}
} //General_Object_Extensions
public static class General_functions
{
public static string fn_ComputerName(string IP)
{
//USAGE
//From: http://stackoverflow.com/questions/1444592/determine-clients-computer-name
//string IP = Request.UserHostName;
//string compName = CompNameHelper.DetermineCompName(IP);
IPAddress myIP = IPAddress.Parse(IP);
IPHostEntry GetIPHost = Dns.GetHostEntry(myIP);
List<string> compName = GetIPHost.HostName.ToString().Split('.').ToList();
return compName.First();
}
static public string fn_ReturnPortionOfStringAfterLastOccuranceOfACharacter(string strInput, char cBreakCharacter)
{
// NOTE: for path backslash "/", set cBreakCharacter = '\\'
string strX = null;
//1] how long is the string
int iStrLenth = strInput.Length;
//2] How far from the end does the last occurance of the character occur
int iLenthFromTheLeftOfTheLastOccurance = strInput.LastIndexOf(cBreakCharacter);
int iLenthFromTheRightToUse = 0;
iLenthFromTheRightToUse = iStrLenth - iLenthFromTheLeftOfTheLastOccurance;
//3] Get the Portion of the string, that occurs after the last occurance of the character
strX = fn_ReturnLastXLettersOfString(iLenthFromTheRightToUse, strInput);
return strX;
}
static private string fn_ReturnLastXLettersOfString(int iNoOfCharToReturn, string strIn)
{
int iLenth = 0;
string strX = null;
int iNoOfCharacters = iNoOfCharToReturn;
iLenth = strIn.Length;
if (iLenth >= iNoOfCharacters)
{
strX = strIn.Substring(iLenth - iNoOfCharacters + 1);
}
else
{
strX = strIn;
}
return strX;
}
} //General_functions
}
I would agree that Log4Net and NLog seem to be the two most commonly used products on the different projects I have been a member.
If you are looking for a great tool that you can use for logging, error handling and anything else where AOP would be beneficial I would highly recommend PostSharp (http://www.postsharp.net/). You set your logging/error handling up centrally and then just decorate methods. It is a well documented and supported product. They have a community license, which is free - and it is free for individuals. They also have professional and ultimate versions of the products, which would make more sense if you're using it as a team.
I don't work at PostSharp :-) I've just used it in the past and really like it.
I searched a lot and still couldn't find a solid solution for this. Suppose you have methods in your application. This methods use "System.Web.Configuration.WebConfigurationManager.OpenWebConfiguration" to access some setting in the web.config. If you try to test these methods, your tests will fail because your test project doesn't have web.config.
What is the best way to solve this problem. For projects with simple config file, I usually use a method like this as facade method.
public class Config
{
public static String getKeyValue(String keyName)
{
if (keyName == String.Empty) return String.Empty;
String result = "";
System.Configuration.Configuration rootWebConfig1 =
System.Web.Configuration.WebConfigurationManager.OpenWebConfiguration(null);
if (rootWebConfig1.AppSettings.Settings.Count > 0)
{
System.Configuration.KeyValueConfigurationElement reportEngineKey =
rootWebConfig1.AppSettings.Settings[keyName];
if (reportEngineKey != null)
{
result = reportEngineKey.Value;
}
}
return result;
}
}
Every time I tried to set the path for OpenWebConfiguration( ), I got the error "The relative virtual path is not allowed"
To make that scenario more testable, I usually take the approach of making a "settings manager" of my own, and giving it an interface. So for example:
public interface IConfig
{
string GetSettingValue(string settingName);
}
Then I can have my "real" implementation:
public sealed class Config : IConfig
{
public string GetSettingValue(string settingName)
{
// your code from your getKeyValue() method would go here
}
}
Then my code that uses it would take in an instance of this (this is an example of the Dependency Inversion Principal):
public void DoStuff(IConfig configuration)
{
string someSetting = configuration.GetSettingValue("ThatThingINeed");
// use setting...
}
So now for my production code, I can call DoStuff and pass in an instance of Config.
When I need to test, I can use a mocking tool (Moq, JustMock, RhinoMocks, etc) to create a fake IConfig that returns a known value without hitting the actual .config file, or you can do it without a mocking framework by making your own mocks (and store them in your test project).
public class ConfigMock : IConfig
{
private Dictionary<string, string> settings;
public void SetSettingValue(string settingName, string value)
{
settings[settingName] = value;
}
public string GetSettingValue(string settingName)
{
return settings[settingName];
}
}
and
[Test]
public void SomeExampleTest()
{
var config = new ConfigMock();
config.SetSettingValue("MySetting", "SomeValue");
var underTest = new MyClass();
underTest.DoStuff(config);
}
The easiest way to do this is to use a mocking library such as moq. It takes a bit of time to figure it out, but once you do you can abstract away most of your plumbing to return the values you need for repeatable, consistent testing.
We are trying to implement MessagePack on our Web API. However I am having trouble finding a reliable up-to-date formatter. I found the WebApiContrib.Formatting.MsgPack formatter but the last update was in 2014 and when I try using it it throws "Method not found" exception.
I also looked at an example here but the code contains some method references that are no longer present in the nuget package.
There seems to be an updated ASP.NET Core library with a set of formatters but we are not looking to move to Core any time soon.
Does someone have a tip for a reliable formatter for MessagePack that would do the trick?
Implementing custom Media Type Formatter is pretty easy task if you have all serialization/deserialization code. So if you can't find suitable solution, you could implement your own. Here is working sample:
public class MessagePackFormatter : MediaTypeFormatter
{
public MessagePackFormatter()
{
SupportedMediaTypes.Add(new MediaTypeHeaderValue("application/x-msgpack"));
}
public override bool CanReadType(Type type)
{
return true;
}
public override bool CanWriteType(Type type)
{
return true;
}
public override Task WriteToStreamAsync(Type type, object value, Stream writeStream, HttpContent content, TransportContext transportContext)
{
if (type == null)
{
throw new ArgumentNullException(nameof(type));
}
if (writeStream == null)
{
throw new ArgumentNullException(nameof(writeStream));
}
MessagePackSerializer.NonGeneric.Serialize(type, writeStream, value, ContractlessStandardResolver.Instance);
return Task.FromResult(0);
}
public override Task<object> ReadFromStreamAsync(Type type, Stream readStream, HttpContent content, IFormatterLogger formatterLogger)
{
if (type == null)
{
throw new ArgumentNullException(nameof(type));
}
if (readStream == null)
{
throw new ArgumentNullException(nameof(readStream));
}
var value = MessagePackSerializer.NonGeneric.Deserialize(type, readStream, ContractlessStandardResolver.Instance);
return Task.FromResult(value);
}
}
As you see it's very thin and just invokes MessagePackSerializer.NonGeneric class for serialization/deserialization. I'd suggest starting using such simple formatter and if you encounter some problems with it for specific cases, just tune the code to fix the issues. You could also examine existing implementation for .Net Core for handling some corner cases.