CodePrairie .NET

South Dakota .NET User Group


August 2007 - Posts

  • Mocking server objects from unit tests

    So heres the problem statement:

    I'd like to be able to use WatiN tests to test my MonoRail views. But the controllers that render these views depend heavily on calling an external webservice that is large (in terms of number of methods and amount of data returned).

    Now I could setup a stub service and have my controllers call that, however I think that keeping track of the data it is supposed to return and when would be very difficult. If only I could create mock instances from my unit tests.

    Hmm, maybe I can, take a look at this nugget:

    MockRepository mocks = new MockRepository();
    AccountControllerImpl mockController = mocks.PartialMock<AccountControllerImpl>();
    using(WebServiceHost host = new WebServiceHost(mockController,new Uri("http://localhost:8901"))) {
     host.AddServiceEndpoint(typeof(IAccountController),new WebHttpBinding(), "Account.svc");
     using (mocks.Record())
      Expect.Call(mockController.Info("")).IgnoreArguments().Return(new XDocument(new XElement("Hello")));
     using (mocks.Playback())
      WatiN.Core.IE ie = new WatiN.Core.IE();

    I then tell my page I'm testing to hit the webservice at a different URL and viola it hits my mock service. To make this really usable I'd probably need to extend RhinoMocks to correctly generate the mock using only the interface (right now I have to use the Implementation class and mark it as InstanceContextMode = Single and have virtual methods)


    Posted Aug 17 2007, 03:49 PM by chrisortman with no comments
    Filed under: ,
  • An Inconvenient Print Job

    Missy: "Could you please print off these 2 recipies and bring them home"

    Chris: "Sure, happy to"

    Chris: "Oh wait...someone is printing 3000 (not exagerating) pages of emails and support incidents. I guess after I refill the tray, twice, so that this crap will finish printing maybe I can schedule some printer time today."


    I wonder if I could get this stopped if I called the EPA?

  • Connecting to SQL Server 2005 with Ruby Sequel

    So here I was wanting to query some database tables that get populated during our build process and then modify some text files / source control based on what's in the table. Seemed like a good excuse to use some ruby but low and behold ActiveRecord doesn't seem to care for the structure of our database tables.

    Enter Sequel

    Sequel is an ORM framework for Ruby. Sequel provides thread safety, connection pooling, and a concise DSL for constructing queries and table schemas.

    Sequel makes it easy to deal with multiple records without having to break your teeth on SQL.

    cool, gem install sequel

    But how to connect to SQL server? Sequel supports DBI, but I could find no tailor made examples (plenty of SQLLite, MySql and Postgres though).

    Well first things first, you'd better make sure you have DBI installed and configured correctly. If you are running using the Windows One Click installer version of ruby you are almost there, you just need to copy the ADO driver. For instructions on that see this rails howto

    I however am running under cygwin and installed the ruby cygwin package so I need to install DBI myself.

    Download the ruby-dbi source from

    Extract your sources (tar -xvf dbi-0.1.1.tar.gz)
    cd to your new folder and do these commands
    ruby setup.rb config --with=dbi,dbd_msql,dbd_ado (don't use spaces after the commas it seems picky)
    ruby setup.rb setup
    ruby setup.rb install

    At this point you might want to jump into irb and make sure all is well with the world.
    require 'dbi' should return true.

    Ok, now let's connect to the database.
    I started by working from this post:

    Sequel version 0.1.3 has just been released. This release adds a DBI
    adapter, and thereby support for ODBC, ADO, Frontbase and other
    databases. For example:

    require 'sequel/dbi'
    DB = 'dbi:/odbc:my_dsn'
    DB[:my_table].all #=> returns all rows in the table as hashes
    # etc...

    As well as this:

    require 'dbi'
    class Server
     attr_reader :name
     def initialize(name)
      @dbh=DBI.connect("DBI:ADO:Provider=SQLNCLI;Data Source=#{name};Integrated Security=SSPI")
     def databases
      @dbh.select_all('SELECT name FROM master.sys.databases ORDER BY 1') do | row |
    class Database
     attr_reader :name
     def initialize(dbh,name)
    server.databases.each {|x| p}

    My first feeble attempt when something like this:

    require 'sequel/dbi'
    db = 'dbi:/ado:Provider=SQLNCLI;Data Source=(local)\\sqlexpress;Integrated Security=SSPI'

    Which failed with URI::InvalidURIError because wants to do some URI parsing and I did not give it a valid URI. Thankfully though I can look at source code.

    My first check is in dbi.rb. If your playing along at home it will be in /usr/lib/ruby/gems/1.8/gems/sequel- on cygwin and C:\ruby\lib\ruby\gems\1.8\gems\sequel-\lib\sequel on windows

    The connect method of this class looks interesting

     def connect
            dbname = @opts[:database] =~ /^DBI:/ ? \
              @opts[:database] : @opts[:database] = 'DBI:' + @opts[:database]
            ::DBI.connect(dbname, @opts[:user], @opts[:password])

    So all it really cares about is a value in an @opts has for :database. So to see how I can get that there I need to look at where @opts comes from. Well I'll save you the trouble and tell you it comes from self.connect in database.rb (connect is aliased to open btw) Have a look at this snippet:

    # Converts a uri to an options hash. These options are then passed
        # to a newly created database object.
        def self.uri_to_options(uri)
            :user => uri.user,
            :password => uri.password,
            :host =>,
            :port => uri.port,
            :database => (uri.path =~ /\/(.*)/) && ($1)
        # call-seq:
        #   Sequel::Database.connect(conn_string)
        #   Sequel.connect(conn_string)
        # Creates a new database object based on the supplied connection string.
        # The specified scheme determines the database class used, and the rest
        # of the string specifies the connection options. For example:
        #   DB = 'sqlite:///blog.db'
        def self.connect(conn_string, more_opts = nil)
          uri = URI.parse(conn_string)
          c = @@adapters[uri.scheme.to_sym]
          raise SequelError, "Invalid database scheme" unless c
 || {}))

    Looking at this it seems like connect and hence open will take an optional second hash that is merged with what gets parsed from the URI. This is promising because I had seen several examples that just passed a scheme to open so I thought I'd try this

    require 'sequel/dbi'
    db = 'dbi:/ado', {:database => 'DBI:ADO:Provider=SQLNCLI;' +
                                              'Data Source=(local)\\sqlexpress;' +
                                              'Integrated Security=SSPI'}

    Hey, that didn't give me an error, in fact it gave me a Sequel::DBI::Database object. Surely it won't work, I'd better check.

    db.execute("SELECT Name from master.sys.databases order by 1").fetch_all

    Well look at that it did work. And there you have it. You've just connected to SQLExpress 2005 with sequel. Happy hunting.

    Posted Aug 10 2007, 06:37 AM by chrisortman with 2 comment(s)
    Filed under:
Powered by Community Server (Commercial Edition), by Telligent Systems