Simon on Software

Running node.js on windows now is not that difficult. I cheated and installed some binaries, rather than building from source via cygwin.

Here are the steps I took:

• Download the binaries from http://node-js.prcn.co.cc/. Find the latest “complete” release which was 0.3.1_2 as of writing.
• Unzip to whereever you like on your disk
• Start cygwin and add node.exe to your path. On my system, I edit /home/smwhit/.bashrc and add the following line at the end export PATH=.:/cygdrive/c/dev/node/bin:`printenv PATH`. Note how you refer to your windows C drive as /cygdrive/[drive letter].
• Follow the Hello World article at http://howtonode.org/hello-node

Hope that helps someone else get started

Just a quick one, as I just lost an hour on this problem.

If you are trying to set up your environment to run with the above environment, you will likely see the following error message when you try and run your app:

Could not create the driver from NHibernate.Driver.SQLite20Driver, NHibernate, Version=2.1.2.4000, Culture=neutral, PublicKeyToken=aa95f207798dfdb4.

and deeper in the stack trace:

The IDbCommand and IDbConnection implementation in the assembly System.Data.SQLite could not be found. Ensure that the assembly System.Data.SQLite is located in the application directory or in the Global Assembly Cache. If the assembly is in the GAC, use element in the application configuration file to specify the full name of the assembly.

If like me, you really had added a reference to System.Data.Sqlite.dll, the you need to add the following to your app.config file:

One of the nice things about open source software is that sometimes you don’t need to write code you think you need to.

I have found that it is possible to generate SpecFlow execution reports with assemblies built against xunit even though SpecFlow doesn’t explicitly support this.

The suprise is that xunit.console will generate nunit compatible Xml Results.

So all we need is –

• xunit.console Bowling.Specflow.dll /nunit TestResult.xml
• specflow nunitexecutionreport ..\..\Bowling.Specflow.csproj

Hope that helps someone.

I struggled yesterday with saving modified entities back to the database after they were disconnected from their original DataContext. After much googling, I found the way forward is to ensure you have a version column in your database with a timestamp datatype, and map this with a Binary field in your mapping metadata. Then you can call an overload of Attach.

A complete working example below, as simple as I can make it. Hope that helps somebody.

using System.Data.Linq;
using System.Data.Linq.Mapping;
using System.Linq;
using NUnit.Framework;

namespace CustomerSaver
{
    [TestFixture]
    public class Class1
    {
        [Test]
        public void TrySaveCustomer()
        {
            const string connectionString = @"server=.\SqlExpress; database=test;trusted_connection=true";
            Customer customer;

            using(DataContext readContext = new DataContext(connectionString))
            {
                customer = readContext.GetTable().First();
            }

            customer.Name = customer.Name + ".";
            using (DataContext saveContext = new DataContext(connectionString))
            {
                saveContext.GetTable().Attach( customer, true );
                saveContext.SubmitChanges();
            }
        }
    }

    [Table(Name="Customers")]
    public class Customer
    {
        [Column(IsPrimaryKey = true)]
        public int Id { get; set; }

        [Column]
        public string Name { get; set; }

        [Column(IsVersion = true, IsDbGenerated = true)]
        public Binary Version { get; set; } //timestamp column in sql server
    }
}

If you are behind a http proxy, and wanting your application to connect to a REST or Web Service, it is likely you will not be able to, and will get a Http Status code of 407 back.

However, you can pop the following into app.config/web.config, and your application will use the same proxy settings as Internet Explorer.

Hope that helps someone else.

Earlier, I mentioned how I was creating an archive copy of a workbook.

I found a more efficient way of doing things that works a lot faster, and is a whole lot less code

Sub SaveFileWithoutFormulas()
  
    If SaveWorkbook Then       
        Dim wsCounter As Long
        Dim ws As Worksheet
        
        For wsCounter = 1 To ThisWorkbook.Worksheets.Count
            Set ws = Nothing
            Set ws = Sheets(wsCounter)
            Sheets(wsCounter).Select
            
            If ws.AutoFilterMode Then
                Application.ActiveSheet.ShowAllData
            End If
            
            Cells.Select
            Selection.Copy
            Selection.PasteSpecial Paste:=xlPasteValuesAndNumberFormats, Operation:= _
                xlNone, SkipBlanks:=False, Transpose:=False
        Next wsCounter    
    End If    
End Sub

Function SaveWorkbook()
    On Error GoTo ErrHandler:
    Dim workBookName As String
    workBookName = Replace(ThisWorkbook.Name, ".xls", "_")
    workBookName = workBookName & Format(Date, "ddmmmmyyy") & ".xls"
    
    ActiveWorkbook.SaveAs Filename:= _
        workBookName _
        , FileFormat:=xlNormal, Password:="", WriteResPassword:="", _
        ReadOnlyRecommended:=False, CreateBackup:=False
    SaveWorkbook = True
    Exit Function
    
ErrHandler:
    SaveWorkbook = False
End Function

Well, I feeling slightly pleased with myself. Last time, I wrote that I had found it difficult to discover how outputting reports in SpecFlow worked. Well, following Ayende’s motto of “I completely agree, can you send me a patch”, I did just that.

I have had my pull request on github accepted, and changes providing some extra documentation are now in the master SpecFlow branch.

It took me a while to figure out how to generate reports in SpecFlow. After all, now we have all these lovely Specification written and the code written to satisfy our requirements, wouldn’t it be nice to generate a report for Project Managers/Customers/Whoever so they can see progress on certain areas of functionality.
There is a command line tool, specflow.exe that can generate two types of reports. The documentation could do with some work, as it is not clear what some of the parameters refer to. I had to look at the source code to find what some parameters meant and what the defaults were. (I am planning on submitting a patch to the team with improved help docs!)

It is useful to put the path for specflow into your path to allow. By default, the installation directoy is C:\Program Files\TechTalk\SpecFlow

Test Execution Report

This is a high level summary showing how many scenarios have passed or failed, as well as timings of the steps. This report uses the NUnit Results XML file as it’s input which wasn’t obvious.

The syntax for this command is:

specflow nunitexecutionreport VisualStudioProjectFile /xmlTestResult:PathToNunitTestResultXMLFile

In my case, I execute:

specflow nunitexecutionreport Specs.csproj /xmlTestResult:bin\debug\TestResult.xml

Step Definition Report

This seems to be a way of seeing what steps you have written, and how many times they are called. We were able to identify a number of steps that had become obsolete during the refactoring of some specs.

The syntax for this command is:

specflow stepdefinitionreport VisualStudioProjectFile

In my case, I execute:

specflow StepDefinitionReport Specs.csproj

This week, I have had a bit of an epiphany with regard to how we should approach automated testing our projects.

Misgivings about TDD style testing

What do I want from automated testing?

The result is a series of BDD styles specs using a .Net project called SpecFlow, which are written in “english” using the gherkin DSL which comes from the Ruby Project Cucumber.

I’m not going to describe the rationale behind BDD as others have done so before, better than I can, here and here

SpecFlow is currently at version 1.2.0 and has versions for Visual Studio 2008 and 2010. The reason for the different versions is that there are specific Item Templates within Visual Studio to aid the creation of SpecFlow Features and Steps.

Steve Sanderson has an excellent post which covers SpecFlow with ASP.net MVC which was my inspiration for exploring this framework in the first place. We have a Windows Forms client, so our Specs cover the layers immediately below the UI (presenter and downwards).

References
SpecFlow Home Page
Cucumber Project
Steve Sanderson Article