Archive

Posts Tagged ‘TDD’

Configuration DSL – Part III: Writing a Fluent Interface for Configuration in C#

July 12th, 2009 No comments

This is the third entry on creating configuration DSLs in C#. The previous entry looked at the implementation. The basis of this code is to solve the problem that there are multiple configurations an application may require: development, test and moving through environments up to production. This solution makes the distinction that there are profiles of settings which are distinct from the values in each setting. Before I address the problems in this design let’s try and see what patterns are used in the DSL and relate them to Fowler’s DSL patterns.

Patterns used …

Problems: Making the configuration immutable behind a factory

Configuration DSL – Part II: Writing a Fluent Interface for Configuration in C#

June 25th, 2009 No comments

This is the second entry on creating configuration DSLs in C#. The previous entry looked at the design of the interface. This one looks at the implementation. Just a small note that the configurator does not use an IoC container to solve the problem of dependency injection.

Separating out settings from values: the configurator and the configuration

This design has two main parts to the implementation. The Configurator and the configuration:

  1. Configurator: this wraps the different types of settings and combinations available
  2. Configuration: these are the actual values for a setting

It is important that there is this distinction. The configurator allows you to set defaults or set values based on other values. In many values this allows you to code implicit understandings – or even profiles – of the system.

For example, in the previous entry I talked about three types of profiles: production, acceptance/integration testing and unit testing. The production profile required logging, proxies and credentials to external systems. The tests profiles required stubs and mocks. Now for some code.

Structure of the project

Here’s a quick overview of the code for the basic test application with configuration. If you want a copy of src there’s one at http://github.com/toddb/DSL-Configuration-Sample. The main point is that there is a folder configuration within the core project. In this case, there is an application that accepts the configuration.

  \Core
    Application.cs
    ApplicationRunner.cs
    IApplication.cs
    IAssetService.cs
    IConnector.cs
    IHttpService.cs
    LogService.cs
    \Configuration
      AppConfiguration.cs
      AppConfigurator.cs
      Credentials.cs
      IAppConfiguration.cs
      IAppConfigurator.cs
      KnownDotNetConfig.cs
      Proxy.cs
  \Models
    Asset.cs
  \Tests
    \Acceptance
      ApplicationRunnerTest.cs
      IndividualUpdateVisualisationTests.cs
    \Configuration
      ConfigurationTest.cs
    \Core
      ApplicationRunnerTest.cs
      ApplicationTest.cs
    App.config

The configuration

The configuration class is as you would expect. It requires a number of values. In this case, it just has public getter/setters on each of them. At this stage, the idea is that any profile of settings requires all of these in some form or rather.

using System.Net;
using Core;
using log4net;

namespace Core.Configuration
{
    public class AppConfiguration : IAppConfiguration
    {
        public IAssetService AssetService { get; set;}
        public IConnector IncomingConnector { get; set; }
        public IConnector OutgoingConnector { get; set; }
        public NetworkCredential OutgoingCredentials { get; set; }
        public WebProxy Proxy { get; set;}
        public string BaseUrl { get; set; }
        public ILog Logger { get; set; }
    }
}

The configurator

The configurator is more complex so let’s look at the interface first. It is the list of items that we saw in the first entry. It is important that this interface is as expressive as possible for helping setting up a profile of settings. In this case, you might note that only have one method each for the incoming and outgoing connectors rather than defaults. That is because I would expect people to ignore setting this up explicitly unless they require it. This is implicit knowledge that may be better handled in other ways. There are some others I’ll leave for now.

  using System;
  using Core;
  using log4net;

  namespace Core.Configuration
  {
      public interface IAppConfigurator : IDisposable
      {
          void IncomingConnector(IConnector connector);
          void OutgoingConnector(IConnector connector);
          void BaseUrl(string url);
          void BaseUrlFromDotNetConfig();
          void UseCredentials(string username, string password);
          void UseCredentialsFromDotNetConfig();
          void RunWithNoProxy();
          void RunWithProxyFromDotNetConfig();
          void RunWithProxyAs(string username, string password, string domain, string url);
          void UseLog4Net();
          void UseLoggerCustom(ILog logService);
      }
  }

Now that you have seen the interface and hopefully have that in your head, now we’ll look at the implementation because this is the guts of the DSL. Don’t be surprised when you see that there really isn’t anything to it. All of the members which are implementing the interface in this case merely update the configuration. Have a read and I’ll explain the main simple twist after the code.

  using System;
  using System.Configuration;
  using System.Net;
  using log4net;

  namespace Core.Configuration
  {
      public class AppConfigurator : IAppConfigurator
      {
          private NetworkCredential _credentials;
          private WebProxy _proxy;
          private string _url;
          private ILog _logger;
          private IConnector _incoming;
          private IConnector _outgoing;

          AppConfiguration Create()
          {
              var cfg = new AppConfiguration
                            {
                                IncomingConnector = _incoming,
                                OutgoingConnector = _outgoing,
                                Proxy = _proxy,
                                OutgoingCredentials = _credentials,
                                BaseUrl = _url,
                                Logger = _logger
                            };
              return cfg;
          }

          public static AppConfiguration New(Action<IAppConfigurator> action)
          {
              using (var configurator = new AppConfigurator())
              {
                  action(configurator);
                  return configurator.Create();
              }
          }

          public void IncomingConnector(IConnector connector)
          {
              _incoming = connector;
          }

          public void OutgoingConnector(IConnector connector)
          {
              _outgoing = connector;
          }

          public void BaseUrl(string url)
          {
              _url = url;
          }

          public void BaseUrlFromDotNetConfig()
          {
              _url = ConfigurationManager.AppSettings[KnownDotNetConfig.BaseUrl];
          }

          public void UseCredentials(string username, string password)
          {
              _credentials = Credentials.Custom(username, password);
          }

          public void UseCredentialsFromDotNetConfig()
          {
              _credentials = Credentials.DotNetConfig;
          }

          public void RunWithNoProxy()
          {
              _proxy = null;
          }

          public void RunWithProxyFromDotNetConfig()
          {
              _proxy = Proxy.DotNetConfig;
          }

          public void RunWithProxyAs(string username, string password, string domain, string url)
          {
              _proxy = Proxy.Custom(username, password, domain, url);
          }

          public void UseLog4Net()
          {
              _logger = LogService.log;
          }

          public void UseLoggerCustom(ILog logService)
          {
              _logger = logService;
          }

          public void Dispose()
          {

          }
      }
  }

All of that is pretty simple. The basis of the interface is to allow the profile either get settings for the App.Config with the methods suffixed with FromDotNetConfig or pass in their own (eg (username, password) or logService). The code in those classes is also straightforward.

Classes loading from the DotNet Config

This is very straightforward class which just gets the value through the ConfigurationManager or passes through the username and password. This implementation isn’t complex – so you may want to not call ConfigurationManager every time. I’ll leave that implementation to you.

  using System.Configuration;
  using System.Net;

  namespace Core.Configuration
  {
      public class Credentials
      {
          public static NetworkCredential DotNetConfig
          {
              get
              {
                  return new NetworkCredential(ConfigurationManager.AppSettings["Asset.Username"],
                                               ConfigurationManager.AppSettings["Asset.Password"]);
              }
          }

          public static NetworkCredential Custom(string username, string password)
          {
              return new NetworkCredential(username, password);
          }

      }
  }

The same approach is used in the Credential class. In this example, it only returns a webproxy if a url was given in the config. Again, here’s another example of implicit profile settings.

  using System;
  using System.Configuration;
  using System.Net;

  namespace Core.Configuration
  {
      public class Proxy
      {

          public static WebProxy DotNetConfig
          {
              get
              {
                  WebProxy proxy = null;
                  if (!String.IsNullOrEmpty(ConfigurationManager.AppSettings["Proxy.Url"]))
                  {
                      proxy = new WebProxy(ConfigurationManager.AppSettings["Proxy.Url"], true)
                                  {
                                      Credentials = new NetworkCredential(
                                          ConfigurationManager.AppSettings["Proxy.Username"],
                                          ConfigurationManager.AppSettings["Proxy.Password"],
                                          ConfigurationManager.AppSettings["Proxy.Domain"])
                                  };
                  }
                  return proxy;
              }
          }
          public static WebProxy Custom(string username, string password, string domain, string url)
          {
              return new WebProxy(username, true)
                         {
                             Credentials = new NetworkCredential(username, password, domain)
                         };
          }

      }
  }

By now you should have seen all the basic getter/setters for values. How now are profiles of settings done?

Consuming the configurator

The configurator initiation code is also pretty straightforward and this implementation uses C# actions. Let’s step through the code outline above in two methods Create and New.

When you call the configurator, if you remember, it was through AppConfigurator.New(configuration). This returns a configuration. This is a simple pattern of calling a static method that returns an class instance. In this, New creates a new instance of itself which on the method Create returns a configuration. So where is the real work done?

  AppConfiguration Create()
   {
       var cfg = new AppConfiguration
                     {
                         IncomingConnector = _incoming,
                         OutgoingConnector = _outgoing,
                         Proxy = _proxy,
                         OutgoingCredentials = _credentials,
                         BaseUrl = _url,
                         Logger = _logger
                     };
       return cfg;
   }

   public static AppConfiguration New(Action<IAppConfigurator> action)
   {
       using (var configurator = new AppConfigurator())
       {
           action(configurator);
           return configurator.Create();
       }
   }



The real work in done in the one line action(configurator). This line calls the action/lambda that you passed in to create values in the configurator and then when Create is called these are merged and returned.

So, back to the original code (see below) the action(configurator) will take each cfg.* and run the method. In this case, the first with run BaseUrlFromDotNetConfig. If that doesn’t make sense, think of it this way, at the point of action in the code, it executes each of the cfg.* lines in the context of that class.

  AppConfigurator.New(cfg =>
                          {
                              cfg.BaseUrlFromDotNetConfig();
                              cfg.RunWithProxyFromDotNetConfig();
                              cfg.UseCredentialsFromDotNetConfig();
                              cfg.UseLog4Net();
                          });

So that’s how to get the configuration available for the application start up. You can get a lot more complex but that should get you started. Personally, if you like this approach then go head off in the direction of libraries doing this: nBehave, NHibernate Fluent interface, StoryQ, codecampserver.

Good luck.

The next part tries to summarise this design, how it relates to Fowler’s patterns in DSLs and some problems with this DSL and solutions to this.http://github.com/toddb/DSL-Configuration-Sample/tree/master

DSL – Part I: Writing a Fluent Interface for Configuration in C#

June 24th, 2009 No comments

I spent the weekend writing a configuration domain specific language (DSL) for a client interface. This interface was lifted from TopShelf project on codeplex after listening to podcast on MassTransit. It was all pretty straightforward and expected as you would imagine. I want to document for myself the process of setting one up for the future and in the process map the code to the patterns that Fowler is outlining in his new DSL book. If you are wondering Fowler would call this an internal DSL or fluent interface. By the end of this, I should have got through where I did some method chaining, used a strategy pattern and how I think it has helped my in DDD. What I find though is that the configurator code is not simply about writing cleaner, prettier code.

Most importantly, I want to think through how this approach is good for TDD. What I have noticed the most is that by using a fluent interface for the configuration I found that it (a) showed up code smells because I tried to consume code more often and in different ways and (b) helped me refactor other parts of the code because they no longer had dependencies on configuration code.

The problem domain …

My application is minor application living in a windows service. Every five minutes, it polls a webservice to get a list of items that it must return information about. That external system has a key for the item and relates that key to the key for the same item in the internal system. This is a trivial system that then gets information from the internal system, collates it and then sends that back to the external system. The external caches this information and combines it with its own information. Quite simple – not if you looked at the early versions of the code.

For this explanation, the system needs to be configured with incoming and outgoing feeds which in turn need network credentials and proxy settings. Both feeds have the same domain, credentials and proxy settings. Early version worked well until you needed to test them – or extend. What I would also like to do is have tests around logging because this system must prove that when it isn’t receiving data that someone is getting notified (via logging framework)

What I also require are acceptance tests, integration tests and unit tests. To run each of these it can be hard if you can’t turn settings on and off. Unfortunately, Microsoft’s app.config isn’t really that good at handling all these situations. To be fair it is only XML. But I don’t see easy solutions to this and as such I find others and myself manually changing the XML at times to test edge cases. That might be fine at coding time but this leaves little opportunity for automated regression testing. My success criteria was to allow for default values (eg corporate proxy is http:/proxy.mycompany.com), default settings (proxy=on) and custom versions of both (eg no proxy or mock proxy) within the same test run.

The simplicity of the solution: using blocks for configuration

Let’s start with my first consumer code. This is from the code that creates the configuration for the application. This code is pretty simple, although looking at here, it also looks a little wordy. Hopefully you read it as I do. With the new configuration, I’ll have the base url with values from the dotnet config, I’ll run app with a proxy with the values in the dotnet config, I’ll use the network credential values from the dotnet config and finally I’ll use log4net. In short, that’s the configuration for a corporate network going out through a proxy to a external address which is secure behind credentials.

  var configuration = AppConfigurator.New(cfg =>
                          {
                              cfg.BaseUrlFromDotNetConfig();
                              cfg.RunWithProxyFromDotNetConfig();
                              cfg.UseCredentialsFromDotNetConfig();
                              cfg.UseLog4Net();
                          });

Just a quick note. For those intellisense-ers out there. Yes, when you hit control-space after cfg dot, you see a list of the configuration settings. Here’s a preview of the settings:

  BaseUrl(url)
  BaseUrlFromDotNetConfig
  RunWithProxyAs(username, password)
  RunWithProxyFromDotNetConfig
  UseCredentialsFromDotNetConfig
  UseCredentials(username, password)
  UseLog4Net
  UseCustomLogger
  OutgoingConnector
  IncomingConnector

Interestingly, this particular configuration is rarely run – this is the production configuration (and will require production values). All I tend to need from this configuration for development is a configurable BaseUrl. I’m generally doing work on the local machine, so I’m not going to need a proxy or credentials to the external service. In this case, I won’t tend to use logging yet. But what I will need is the ability to mock out the external service. But before I go there here is how the configuration is hooked into the application. I pass the configuration into the application. [what I will explain later is that at this point I hand in the configuration values having worked out what they are based on the configuration type.]

  TTLApplication.Run(configuration);

Now to the test. These are far more important for development. In the example below, I want to check that in running my application that the Outgoing connector calls the send method (ie that it sends out). In this case, I use Moq as a framework to mock-out the service. This service will is actually just an Http service makes web requests. The code sets up connector to respond to the Send method, runs the application and then verifies that Send was called.

  [TestMethod]
  public void CanMockOutgoingConnector()
  {
      var connector = new Mock<IOutgoingConnector>();
      connector.Setup(x => x.Send());

      Application.Run(AppConfigurator.New(cfg =>
                        { 
                          cfg.OutgoingConnector(connector.Object)
                        }));
      connector.Verify(x => x.Send(), Times.Exactly(1));
  }

You’ll note in both cases, you only setup the what you need. In the first example, the outgoing connector is setup by default. Whereas in the second, it is only the outgoing connector which is setup with a mock.

This configurator settings are in fact more than what you need. The configurator is really good for creating settings – defaults and combinations – but for unit testing it is easier to get down the setting values. Here’s a simple example of above. While there’s not a lot of difference, the important thing is that you are working through less layers because when it gets more complicated you’ll need a greater understanding.

  [TestMethod]
  public void CanMockOutgoingConnector()
  {
      var connector = new Mock<IOutgoingConnector>();
      connector.Setup(x => x.Send());

      Application.Run(new AppConfiguration
                          {
                              IncomingFleet = null,
                              OutgoingUtilization = connector.Object,
                              Url = "http://example.com"
                          });
      connector.Verify(x => x.Send(), Times.Exactly(1));
  } 

At this point, you should be starting to get the idea that this is an easier way to configure the system for multiple purposes. The next entry will look through the code that implements it.

<br />

Fakes Arent Object Mothers Or Test Builders

November 20th, 2008 No comments

Fakes are not Object Mothers or Test Data Builders and Extensions are a bridge to aid BDD through StoryQ

I’ve just spent the week at Agile2008 and have had it pointed out to me that I have been using Fakes as nomenclature when in fact it is an Object Mother pattern. So I have set off to learn about the Object Mother pattern and in doing so came across the Test Data Builder pattern. But because I am currently working in C# 3.5 and using PI/POCOs/DDD, the builder pattern at looking a little heavy and the criticism of Object Mother holds that it ends up as a God object and hence a little too cluttered.

I’ve also found utilising Extensions in C# 3.5 as a good way to keep the OM clean. By adding extensions, it cleans up the code significantly such that BDD through StoryQ became attractive. Extensions have these advantages:

  • Allows you to put a SetUp (Build/Contruct) method on the object itself
  • Makes this/these methods only available within the test project only
  • Keeps data/values separate from setup ie as a cross-cutting concern
  • In tests, specific edge data – or basic tests – aren’t put into object mothers
  • Extensions end up concise and DRY code

Background

I have been using a Fake models as a strategy to unit test (based on Jimmy Nillson’s DDD in C# book). Technically, these are not fakes. According to sites, fakes are a strategy to replace service objects rather than value objects. My use of fakes is rather the “Object Mother” pattern. (Am I the only person six years behind? Probably.) Since knowing about this pattern, I’ve found it on Fowler’s bliki and the original description at XP Universe 2001 (however, I haven’t been able to download the paper as yet – xpuniverse.com isn’t responding).

Having read entries around this pattern, another pattern emerged: the “Test Data Builder” pattern. It is a likable pattern. (In fact, it could be leveraged as an internal DSL - but I haven’t pursued that as yet.) But, given the work I do in C#, it looks a little heavy as it is useful to cope with complex dependencies. In contrast, the fake/object mother I have been using has been really easily to teach, and well liked by, developers.

Basic approach

To test these approaches, I am going to create a simple model: User which has an Address. Both models have validations on the fields and I can ask the model to validate itself. You’ll note that I am using Castle’s validations component. This breaks POCO rules of a single library but in practice is a good trade-off. My basic test heuristic:

  • populate with defaults = valid
  • test each field with an invalid value
  • ensure that inherited objects are also validated

Four Strategies:

  1. Object Mother
  2. Test Data Builder
  3. Extensions with Object Mother
  4. StoryQ with Extensions (and Object Mother)

Strategy One: Object Mother

First, the test loads up a valid user through a static method (or property if you wish). I use the convention “Valid” to always provide back what is always a valid model. This is a good way to demonstrate to new devs what the exemplar object looks like. Interesting, in C# 3.5, the new constructor convention works very well here. You can read down the list of properties easily, often without need to look at the original model code. Moreover, in the original object, there is no need for an overloaded constructor.

[Test]
 public void ValidUser()
 {
     var fake = TestUser.Valid();
     Assert.IsTrue(fake.IsOKToAccept());
 }

  public class TestUser
   {
       public static User Valid()
       {
           return new User {
               Name = "Hone", 
               Email = "Hone@somewhere.com", 
               Password="Hone", 
               Confirm = "Hone"
       }
   }

Oh, and here’s the User model if you are interested. Goto to Castle Validations if this isn’t clear.

 public class User
 {
     [ValidateLength(1, 4)]
     public string Name { get; set; }

     [ValidateEmail]
     public string Email { get; set; }

     [ValidateSameAs("Confirm")]
     [ValidateLength(1,9)]
     public string Password { get; set; }
     public string Confirm { get; set; }

     public bool IsOKToAccept()
     {
         ValidatorRunner runner = new ValidatorRunner(new CachedValidationRegistry());
         return runner.IsValid(this);
     }
  }

Second, the test works through confirming that each validation works. Here we need to work through each of the fields checking for invalid values. With the object mother pattern, each case is a separate static method: email, name and password. The strategy is to always always start with the valid model, make only one change and test. It’s a simple strategy, easy to read and avoids duplication. The problem with this approach is hiding test data; we have to trust that the method, “invalidName”, is actually an invalid name or follow it through to the static method. In practice, it works well enough and avoids duplication.

public static User InvalidEmail()
{
    var fake = Valid();
    fake.Email = "invalid@@some.com";
    return fake;
}

public static User InvalidName()
{
    var fake = Valid();
    fake.Name = "with_more_than_four";
    return fake;
}

public static User InvalidPassword()
{
    var fake = Valid();
    fake.Password = "not_same";
    fake.Confirm = "different";
    return fake;
}

[Test]
public void InValidEmail()
{
    var fake = TestUser.InvalidEmail();
    Assert.IsFalse(fake.IsOKToAccept());
}

[Test]
public void InValidName()
{
    var fake = TestUser.InvalidName();
    Assert.IsFalse(fake.IsOKToAccept());
}

[Test]
public void InValidPassword()
{
    var fake = TestUser.InvalidPassword();
    Assert.IsFalse(fake.IsOKToAccept());
}

Strategy Two: Test Data Builder

I’m not going to spend long on this piece of code because the builder looks like too much work. The pattern is well documented so I assume you already understand it. I do think that it could be useful if you want to create an internal DSL to work through dependencies. Put differently, this example is too simple to demonstrate a case for when to use this pattern (IMHO).

First, here’s the test for the valid user. I found that I injected the valid user behind the scenes (with an object mother) so that I could have a fluent interface with Build(). I can overload the constructor to explicitly inject a user too. I have two options when passing in an object: (1) inject my object mother or (2) inject a locally constructed object. The local construction is useful for explicitly seeing what is being tested. But really, it is the syntactic sugar of C# that is giving visibility rather than the pattern; so, for simple cases, the syntax of the language renders the pattern more verbose.

[Test]
public void ValidUser()
{
    var fake = new UserBuilder().Build();
    Assert.IsTrue(fake.IsOKToAccept());

}

[Test]
public void LocalUser()
{
    var fake = new UserBuilder(TestUser.Valid()).Build();
    Assert.IsTrue(fake.IsOKToAccept());

    fake = new UserBuilder(new User
                    {
                        Name = "Hone", 
                        Email = "good@com.com",
                        Password = "password",
                        Confirm = "password",
                        Address = new Address
                                      {
                                          Street = "Fred",
                                          Number = "19"
                                      }
                    })
                    .Build();
    Assert.IsTrue(fake.IsOKToAccept());
}

public class UserBuilder
{
    private readonly User user;

    public UserBuilder()
    {
        user = TestUser.Valid();
    }

    public UserBuilder(User user)
    {
        this.user = user;
    }

    public User Build() { return user; }
}

Second, let’s validate each field. In the code, and on the positive side, it is clear what constitutes an invalid value and it caters for dependencies such as in “withPassword” between password and confirm. But, really, there is just too much typing; I have to create a method for every field and dependency. For simple models, I am not going to do this. For complex or large models, it would take ages.

[Test]
public void InValidEmail()
{
    var fake = new UserBuilder()
        .withEmail("incorect@@@emai.comc.com.com")
        .Build();
    Assert.IsFalse(fake.IsOKToAccept());
}

[Test]
public void InValidName()
{
    var fake = new UserBuilder()
        .withName("a_name_longer_than_four_characters")
        .Build();
    Assert.IsFalse(fake.IsOKToAccept());
}

[Test]
public void InValidPassword()
{
    var fake = new UserBuilder()
        .withPassword("bad_password")
        .Build();
    Assert.IsFalse(fake.IsOKToAccept());
}

public class UserBuilder
{
    private readonly User user;

    public UserBuilder()
    {
        user = TestUser.Valid();
    }

    public UserBuilder(User user)
    {
        this.user = user;
    }

    public UserBuilder withEmail(string email)
    {
        user.Email = email;
        return this;
    }

    public UserBuilder withPassword(string password)
    {
        user.Confirm = password;
        user.Password = password;
        return this;
    }

    public UserBuilder withName(string name)
    {
        user.Name = name;
        return this;
    }

    public UserBuilder withAddress(Address address)
    {
        user.Address = address;
        return this;
    }

    public User Build() { return user; }
}

Strategy Three: Object Mother with Extensions

Having tried the two previous patterns, I now turn to what extensions have to offer me. (Extensions are something I have been meaning to try for a while.) As it turns out extensions combined with the new constructors allow for SOC and DRY and also allow us to separate valid and invalid data in tests. There is a downside of course. It requires some (reflection) code to make it play nicely. More code – a slippery slope some might say …

First, I have added a new method to my User model in the form of an extension. I have named it SetUp so that it translates into the setup and teardown (init and dispose) phases of unit testing. I could use Build or Construct instead. This method returns my object mother. I still like to keep my object mother data separate because I think of construction and data as separate concerns.

[Test]
 public void ValidUser()
 {
     var fake = new User().SetUp();
     Assert.IsTrue(fake.IsOKToAccept());
 }

 public static User SetUp(this User user)
 {
     User fake = TestUser.Valid();
     return fake; 
 }

I also want to test how to create a version of the user visible from the test code. This is where more code is required to combine the any provided fields from the test with the default, valid model. In returning hydrated test user, this method accepts your partially hydrated object and then adds defaults. The goal is that your unit test code only provides specifics for the test. The rest of the valid fields are opaque to your test; the code reflects on your model only hydrating empty fields so that validations do not fail. This reflection code is used by all SetUp methods across models.

[Test]
 public void LocalUser()
 {
     var fake = new User { Name = "John", Email = "valid@someone.com" }.SetUp();
     Assert.IsTrue(fake.IsOKToAccept());
 }

 public static User SetUp(this User user)
 {
     User fake = TestUser.Valid();
     SyncPropertiesWithDefaults(user, fake);
     return fake; 
 }

 private static void SyncPropertiesWithDefaults(object obj, object @base)
   {
       foreach (PropertyInfo prop in obj.GetType().GetProperties())
       {
           object[] val = new object[1];
           val[0] = obj.GetType().GetProperty(prop.Name).GetValue(obj, null);
           if (val[0] != null)
           {
               obj.GetType().InvokeMember(prop.Name, BindingFlags.SetProperty, Type.DefaultBinder, @base, val);
           }
       }
   }

Second, let’s again look at the code to validate all the fields. Now, note at this point there no other object mothers. I see this strategy says, use object mothers to model significant data and avoid cluttering these with edge cases. The result is to hand in the field/value to isolate and then test. I find this readable and it addresses the concern of not making visible the edge case (invalid) data.

[Test]
public void InValidEmail()
{
    var fake = new User { Email = "BAD@@someone.com" }.SetUp();
    Assert.IsFalse(fake.IsOKToAccept());
}

[Test]
public void InValidName()
{
    var fake = new User { Name = "too_long_a_name" }.SetUp();
    Assert.IsFalse(fake.IsOKToAccept());
}

[Test]
public void InValidPassword()
{
    var fake = new User { Password = "password_one", Confirm = "password_two" }.SetUp();
    Assert.IsFalse(fake.IsOKToAccept());
}

Using extensions combined with constructors are nice. We can also format the page to make it look like a fluent interface if need. For example:

[Test]
public void InValidPassword()
{
    var fake = new User 
                { 
                    Password = "password_one", 
                    Confirm = "password_two" 
                }
                .SetUp();
    Assert.IsFalse(fake.IsOKToAccept());
}

Strategy Four: BDD and StoryQ

Having worked out that using extensions reduce repetitious code, I still think there is a smell here. Are those edge cases going to add value in the current form? They really are bland. Sure, the code is tested. But, really, who wants to read and review those tests? I certainly didn’t use them to develop the model code; they merely assert overtime that my assumptions haven’t changed. Let’s look at how I would have written the same tests using BDD and in particularly StoryQ. Here’s a potential usage of my User model.

Story: Creating and maintaining users

  As an user
  I want to have an account
  So that I can register, login and return to the site

  Scenario 1: Registration Page
    Given I enter my name, address, email and password     
    When username isn't longer than 4 characters           
      And email is validate                                
      And password is correct length and matches confirm   
      And address is valid                                 
    Then I can log now login
      And I am sent a confirmation email

Leaving aside any problems in the workflow (which there are), the story works with the model in context. Here is the code that generates this story:

[Test]
public void Users()
{
    Story story = new Story("Creating and maintaining users");

    story.AsA("user")
        .IWant("to have an account")
        .SoThat("I can register, login and return to the site");

    story.WithScenario("Registration Page")
        .Given(Narrative.Exec("I enter my name, address, email and password", 
                 () => Assert.IsTrue(new User().SetUp().IsOKToAccept())))
        
        .When(Narrative.Exec("username isn't longer than 4 characters", 
                 () => Assert.IsFalse(new User{Name = "Too_long"}.SetUp().IsOKToAccept())))

        .And(Narrative.Exec("email is validate", 
                 () => Assert.IsFalse(new User { Email = "bad_email" }.SetUp().IsOKToAccept())))

        .And(Narrative.Exec("password is correct length and matches confirm", 
                 () => Assert.IsFalse(new User { Password = "one_version", Confirm = "differnt"}.SetUp().IsOKToAccept())))

        .And(Narrative.Exec("address is valid", 
                  () => Assert.IsFalse(new User { Address = new Address{Street = null}}.SetUp().IsOKToAccept())))
        
        .Then(Narrative.Text("I can log now login"))
        .And(Narrative.Text("I am sent a confirmation email"));

    story.Assert();
}

Working with BDD around domain models and validations makes sense to me. I think that it is a good way to report validations back up to the client and the team. There is also a good delineation between the “valid” model (“Given” section) and “invalid” (“When” section) edge cases in the structure. In this case, it also demonstrates that those models so far do nothing because there are no tests in the “Then” section.

Some quick conclusions

  1. Syntactic sugar of new constructor methods in C# 3.5 avoids the need for “with” methods found the Test Data Builder pattern
  2. Extensions may better replace the Build method in the builder
  3. Object mothers are still helpful to retain separation of concerns (eg data from construction)
  4. Go BDD ;-)

Well that’s about it for now. Here’s a download of the full code sample in VS2008.

Postscript: About Naming Conventions

How do I go about naming my object mother? I have lots of options: Test, Fake, ObjectMother, OM, Dummy. Of course, if I was purist it would be ObjectMother. But that doesn’t sit well with me when explaining it others. Although it does make explicit the pattern being used, I find fake most useful (eg FakeUser). It rolls off the tongue, it is self evident enough to outsiders as a strategy. Test (eg TestUser), on the other hand, is generic enough that I have to remind myself of its purpose and then I find I always need to do quick translations from the tests themselves. For example, with UserTest and TestUser I have read them to check which I am working with. For these samples, I have used TestUser. If you come back to my production code you will find FakeUser. I hope that doesn’t prove problematic in the long run as it isn’t really a fake.

jQuery and testing — JSUnit, QUnit, JsSpec [Part 2]

November 15th, 2008 No comments

Trying QUnit with jQuery

In the JSUnit entry, one of the main problems was with the sequencing of calls. Let’s see how QUnit handles this. QUnit has a simple implementation to this problem: stop() and start() commands to synchronise sequences. The basic approach is that it calls a test function. With the stop() command, it will not call the next test until you say start(). But it does complete the rest of the current test.

The code that I ended up with was just what I wanted compared to JsUnit. Previously, I had said that basically I wanted to call my function and then process the result that it was in fact what I wanted. Here is the main javascript code (it’s nice and concise).

One point about the code at this stage. The tests had to be inside the success function to be run. I wonder if this is going to create a code smell in the long run. Plus there is no setup/teardown cycles, again I wonder what that will mean in the long run. Perhaps not?

module("XML to object");

test("Check the xml to object conversion without showing the tree", function() {
  
  expect( 5 )
  stop();
  
   $().storyq({
        url: 'data/results-01.xml', 
        load: '',
        success: function(feed) {
          ok( feed, "is an object" )
          ok( !$.isFunction(feed), "is not a function" )
          ok( feed.version, "has a version: " + feed.version )
          ok( feed.items, "has items")
          same( feed, reference, "is the same as the reference object in data/results-01.js")
          start();
        }
    });

});

Honestly, it is that easy.

Here’s a couple of features that took me half an hour to work out restated from above. (1) stop() and start() almost work as you expect – but I had to put some alerts in to check the order of execution. Basically, a stop() halts any new tests from executing but it keeps the current test executing. The effect of this is that the asynchronise call can be completed. start() then tells the tests to start running again. If you don’t have a start() then you will find that your test runner halts altogether. There is another option and that is to put a timer on the stop and then you don’t need a start. I prefer to keep the tests running as quickly as possible.

Just another note. I decided to do a same() comparison. I have saved and preloaded an object from a file for ease. This kept the test easy to read – my reference object here is quite long. You can see the insertion of this file in the entire code below <script type="text/javascript" src="data/results-01.js"/>

&lt;html>
&lt;head>
  &lt;title>Unit tests for StoryQ Results viewer&lt;/title>
  &lt;link rel="stylesheet" href="../../lib/qunit/testsuite.css" type="text/css" media="screen" />

  &lt;link rel="stylesheet" href="../../lib/treeview/jquery.treeview.css" />
  &lt;link rel="stylesheet" href="../../src/css/storyq.treeview.css" />
  &lt;link rel="stylesheet" href="../../src/css/storyq.screen.css" />

  &lt;script src="../../lib/jquery/jquery.js">&lt;/script>
  &lt;script src="../../lib/jquery/jquery.cookie.js" type="text/javascript">&lt;/script>
  &lt;script src="../../lib/treeview/jquery.treeview.js" type="text/javascript">&lt;/script>
  &lt;script src="../../src/storyq.js" type="text/javascript">&lt;/script>
  &lt;script src="../../src/storyq.treeview.js" type="text/javascript">&lt;/script>
  &lt;script src="../../src/storyq.xml.js" type="text/javascript">&lt;/script>
  &lt;script src="../../src/storyqitem.js" type="text/javascript">&lt;/script>
  &lt;script src="../../src/storyqresults.js" type="text/javascript">&lt;/script> 

  &lt;script type="text/javascript" src="../../lib/qunit/testrunner.js">&lt;/script> 
  &lt;script type="text/javascript" src="data/results-01.js">&lt;/script>  
  
  &lt;script type="text/javascript">
    module("XML");

    test("Check the xml to object conversion without showing the tree", function() {

      expect( 5 )
      stop();

       $().storyq({
            url: 'data/results-01.xml', 
            load: '',
            success: function(feed) {
              ok( feed, "is an object" )
              ok( !$.isFunction(feed), "is not a function" )
              ok( feed.version, "has a version: " + feed.version )
              ok( feed.items, "has items")
              same( feed, reference, "is the same as the refefence object in data/results-01.js")
              start();
            }
        });

    });
  &lt;/script>
  
&lt;/head>
&lt;body>
  
 &lt;h1>QUnit tests&lt;/h1>
 &lt;h2 id="banner">&lt;/h2>
 &lt;h2 id="userAgent">&lt;/h2>
 &lt;ol id="tests">&lt;/ol>

 &lt;div id="main">&lt;/div>

&lt;/div>
&lt;/div>

&lt;/body>
&lt;/html>

The output results from QUnit are nice to look at and easy to read. I didn’t have a couple of errors that weren’t the easiest to debug given the output. Partly though because I was new to it, I was taking too big a step at times!

I’m happy with QUnit – and there are plenty of examples in the JQuery test suite. I can see that I would do TDD with this.

Being a BDD type of guy, I’m now off to see what JsSpec has to offer.

jQuery and testing – JSUnit, QUnit, JsSpec [Part 1]

November 15th, 2008 No comments

Trying JsUnit with JQuery

I have started first with JSUnit because it is tried and true (and to tell the truth I thought it would be fine and didn’t bother with a quick search for alternatives).

For the impatient, I won’t be going with JSUnit and here are some reasons:

  • the problem is that the integration of a setup (ie onload – pausing) to load the data doesn’t integrate well with jQuery. JSUnit has its own setup and document loader but I am still wondering how to do this transparently (ie I didn’t actually get the test to work – I wasn’t exhaustive but then again I don’t think that I should have needed to be to get this test going)
  • Firefox 3.0 on mac doesn’t integrate well (ie it doesn’t work), but safari does! Unfortunately, I am a little bound to firebug for development.
  • JSUnit doesn’t report tests well either

I went and had a look at how JSUnit does it. (Remember that this framework has been around a lot longer than jQuery.) Here is the extract from the code/test samples. The basic setup is to hook into an existing testManager that existing within a frame and then get the data from there. Furthermore, you need to manage your own flag that the process has been complete. JSUnit then looks through all functions that start with test in this case testDocumentGetElementsByTagName checks expected data. Here I assume that the tests are run in a particular frame (buffer()) that testManager gives us access to.

var uri = 'tests/data/data.html';

function setUpPage() {
    setUpPageStatus = 'running';
    top.testManager.documentLoader.callback = setUpPageComplete;
    top.testManager.documentLoader.load(uri);
}

function setUpPageComplete() {
    if (setUpPageStatus == 'running')
        setUpPageStatus = 'complete';
}

function testDocumentGetElementsByTagName() {
    assertEquals(setUpPageStatus, 'complete');
    var buffer = top.testManager.documentLoader.buffer();
    var elms = buffer.document.getElementsByTagName('P');
    assert('getElementsByTagName("P") returned is null', elms != null);
    assert('getElementsByTagName("P") is empty', elms.length > 0);
}

Below is the rewritten code to exercise my code. Here’s a couple of the features:

  • for setup, pass in the correct xml file via uri variable (obviously)
  • to test, I have written a test testXML2Object.

    p. There is one major design problem with the code itself that didn’t allow me to use my own data loader. You will see the line var feed = new StoryQResults(buffer);. Where did that come from? It is nothing close to the code I said that I wanted to exercise. It is infact from within the code I want to exercise. The major issue I found here is that to load and test data I had to use the testManager rather than use my own ().storyq() call.

The other problem was it wouldn’t return the result that I wanted either. I was expecting my feed variable to be an object of the results. Instead I was getting a reference to function StoryQResults – given now that it wasn’t running in Firefox and I didn’t have Firebug life was getting a little hard.

var uri = '../../xml/results-01.xml';

function setUpPage() {
    setUpPageStatus = 'running';
    top.testManager.documentLoader.callback = setUpPageComplete;
    top.testManager.documentLoader.load(uri);
}

function setUpPageComplete() {
    if (setUpPageStatus == 'running')
        setUpPageStatus = 'complete';
}

function testXML2Object() {
    assertEquals(setUpPageStatus, 'complete');
    var buffer = top.testManager.documentLoader.buffer();
    
    var feed = new StoryQResults(buffer);               

    assertEquals(feed.version, '0.1')
    assertEquals("Number of stories", $(feed).size(), 1)
    $.each(feed, function(){
      alert(this)               
    })

}

Even though I know that I getting a function returned instead of an object I am still going to see if I can invoke my own loading function within JSUnit. Here’s what the code would look like below. I wouldn’t recommend running it – just take a look at it. The code to me is a mixtures of styles that start to bloat the code. On the one hand, JSUnit has this setup phase with explicit flags and no anonymous functions. On the other hand, because I am using JQuery conventions, I encapsulate alot of that logic. For example, jQuery(function(){}) waits for the page to be loaded before executing ("#tree").story(), Then I have the callback function inline. It looks good from the outside, but it doesn’t work.

The order of calls is: loading, in test and then loaded. Indicating that my JQuery function runs after the test has been run. The order should have been loading, loaded and then in test. In this sense, while setUpPage runs within its own setup/test/teardown cycle. But my JQuery call isn’t linked into this. JQuery waits is waiting on a document flag rather than a custom flag (within testManager). At this point, I do not wish to dig into these libraries to work them out to get it to all play nicely. It wasn’t designed to work this way. Let’s find one that was.

var data = '';

function setUpPage() {
    setUpPageStatus = 'running';
    alert('loading')
    jQuery(function() {
      $("#tree").storyq({
          url: '../../xml/results-01.xml', 
          success: null, 
          load: function(feed) {
            data = feed;
            alert('loaded')
            setUpPageComplete()
            }
          });
    });
}

function setUpPageComplete() {
    if (setUpPageStatus == 'running')
        setUpPageStatus = 'complete';
}

function testXML2Object() {
    alert('in test')
    assertEquals(setUpPageStatus, 'complete');

    assertEquals(data.version, '0.1')
    assertEquals("Number of stories", $(feed).size(), 1)
}

I’m invoking the two-feet principle: I’m moving onto the next framework (after a quick search): QUnit

jQuery and testing – JSUnit, QUnit, JsSpec [Part 3]

November 14th, 2008 No comments

Trying JsSpec with jQuery

This is part three of three. The previous two have been focussed around managing the problem of timing: JSUnit got too hard and QUnit is easy but you still have to manage timings yourself. With JsSpec there is no problem because it is all managed for you. Nice work! Here’s the code I had to write.

A couple of things to writing it. I had to dig into the code to find out the setup/teardown lifecycle keywords. There turns out to be setup/teardown per test (eg before) and per test suite (eg before all). I also had to dig around to find then comparators (eg should_be, should_not_be_null). I couldn’t find any documentation.

describe('I need to read the xml and convert into object', {
  'before all': function() {
    target = {};
    $().storyq({
        url: 'data/results-01.xml', 
        load: '',
        success: function(feed) {
          target = feed
      }
    })
   
  },
  
  'should return an object': function() {
    value_of(target).should_not_be_null()
  },
  
  'should not be a function': function() {
    value_of(typeof target).should_not_be(typeof Function )
  },
  
  'should have a version': function(){
    value_of(target.version).should_be('0.1')
  },
  
  'should have items': function(){
    value_of(target.items).should_not_be_empty()
  },
  
  'should have the same value as the reference object in data/results-01.js': function(){
    value_of(reference).should_not_be_undefined()
    value_of(target).should_be(reference)
  },
  
})

Ihe output looks nice too ;-) Here’s the overall code. Notice that I have also used the technique of a reference object in results-01.js:

&lt;html>
&lt;head>
&lt;title>JSSpec results&lt;/title>
&lt;link rel="stylesheet" type="text/css" href="../lib/jsspec/JSSpec.css" />
&lt;script type="text/javascript" src="../lib/jsspec/JSSpec.js"/>

&lt;script src="../lib/jquery/jquery.js"/>
&lt;script src="../lib/jquery/jquery.cookie.js" type="text/javascript"/>
&lt;script src="../lib/treeview/jquery.treeview.js" type="text/javascript"/>
&lt;script src="../build/dist/jquery.storyq.js" type="text/javascript"/>

&lt;script type="text/javascript" src="data/results-01.js"/>
&lt;script type="text/javascript" src="specs/treeview.js"/>  

&lt;script type="text/javascript">

  describe('I need to read the xml and convert into object', {
    'before all': function() {
      target = {};
      $().storyq({
          url: 'data/results-01.xml', 
          load: '',
          success: function(feed) {
            target = feed
        }
      })

    },

    'should return an object': function() {
      value_of(target).should_not_be_null()
    },

    'should not be a function': function() {
      value_of(typeof target).should_not_be(typeof Function )
    },

    'should have a version': function(){
      value_of(target.version).should_be('0.1')
    },

    'should have items': function(){
      value_of(target.items).should_not_be_empty()
    },

    'should have the same value as the reference object in data/results-01.js': function(){
      value_of(reference).should_not_be_undefined()
      value_of(target).should_be(reference)
    },

  })
&lt;/script>

&lt;/head>
    &lt;body>
      &lt;div style="display:none;">&lt;p>A&lt;/p>&lt;p>B&lt;/p>&lt;/div>
    &lt;/body>
&lt;/html>

JSSpec isn’t written using JQuery. So there are a couple of issues that I can’t pin down. When I get errors it stops the tests completely. I suspect that this is because these tests are using callbacks and they don’t return an (JQuery) object. JQuery does alot of object chaining and JsSpec isn’t cut out for it (I think).

Well, that’s it.

jQuery and testing – JSUnit, QUnit, JsSpec [Introduction]

November 12th, 2008 No comments

I had been writing a jQuery parser and then realised once I had spiked it that I hadn’t actually written any tests. So, these are some results from a spike in unit testing a jQuery plugin.

Some background, the plugin is a results viewer from an xml feed from storyq. So, I have run some tests and have results. Now I want to see them in html format. The plugin merely transforms the xml to be displayed using the treeview plugin. I wanted to avoid handing in a json feed formatted specifically for the treeview. I wanted all this to happen client side.

The tests have two aspects:

  • xml loading and parsing into an object
  • rendering the object into a tree (at that point treeview takes over)

    In short, I want to test the code underlying this call that returns the feed before populating an <ul id="tree"> element:
$('#tree').storyq({
    url: 'tests/data/results-01.xml',   
    success: function(feed) {

      $("#tree").treeview(); //populate the tree
  
  }    
});

Problem for any framework: sequencing

Let’s take a look at what I want as test code. In this code, I want to populate data with the feed variable returned in the success callback. The test can then check for values. Take a look at the code below. When I run the code, I should (ideally) see the sequence of alerts: loaded, start test, end test. Of course, I don’t. I see start start test, end test, loaded as the sequence. That should be obvious that the callback success hasn’t been called as yet: javascript is run sequentially. Okay, nothing here is surprising. I laboured this point because any of the frameworks must deal with this problem.

var data = {};

jQuery(function() {
  $().storyq({
      url: '../../xml/results-01.xml', 
      success: function(feed) {
        data = feed;
        alert('loaded')       
        }
      });
});

function testXML2Object() {
  alert('start test')
  assertEquals(data, {"result"}, "Assume that result is a real/correct object");
  alert('end test')     
}

Frameworks looked at

I looked at three frameworks for xml loading and parsing:

Conclusions

  • In short, QUnit and JsSpec are both as easy as the other. JSUnit seems now to be too heavy given what we now have.
  • QUnit is a TDD framework is used by jQuery itself. I suspect it might survive longer. There are no setup/teardown phases.
  • JsSpec is a BDD framework and doesn’t use jQuery at all but can easily be used with JQuery plugins. There are good setup/teardown phases for tests and suites.
  • Your choice between the two is likely to be your preference between TDD or BDD. It probably depends upon which boundaries you want to test and how you want to go about specifying.

What I didn’t do:

  • integration with a CI system
  • cross-browser testing
  • test with selenium or watir