C#6, Roslyn and vNext – Bunch of Features

Only a list of various articles about C#6, Roslyn and vNext features…


CT Downloader

CTDownloader is a simple console-based (because of abitity to run under Linux) application for downloading video from in an easy way.
It is possible run under Windows or Linux (for more details see readme.txt in the archive)
Basic usage:

  1. Open Internet Explorer (IE)
  2. Run the application
  3. In IE, go to web with the video
  4. Application will signalize that the stream was found and let you select quality of the stream
  5. Set quality choice
  6. That’s all




Before a few weeks, I was playing with hosting a private NuGet server and creation of own packages. For this purpose I have created a few packages from my small solution called Marazt.Commons and published them on

This package contains base class for WPF development like base class for view-model with various extension method for binding, property notifying, etc.

Contains implementation of some desing patters, e.g. Strategy.

This package contains helper classes for Xml file validation and transformation suing MS XDT. It is used in ConfigTransformation extension.

Contains aspect classes based on PostSharp aspects, e.g. advance default value aspect.

I want to publish it GitHub in the future too.


Generation of Multiple App(web) Config Files in the Solution

In our project we have multiple (seven) configurations for various environments – development, nightly build server, test env. 1, test env. 2, test env. 3, … . For all these environments whe have transformations based on MS Xml Document Transformation. These transformation can be run automatically by VS during deployment or you can use extension SlowCheetah and define transformations in your MSBuild. But this was not suitable for us. We are often switching between environments during the development, so we need to change configuration quickly – for example, just right+click on the transofmation file and run it. There was no existing solution providing these abilities. For this purpose I have created a VS extension ConfigTransfomation. You can specify various transformation files for one configuration file (and you don’t have to create a build definition for them!). Just right-click on it and see the thre preview of the transformation or run it.

Config transformation context menu options

EF Code First & Generate SQL Inserts From Seed Method

Entity Framework Code First is a very popular way of developing ORM based systems these days. Creating and updating Db schema using migrations is quite simple (except situations when ypu are updating types of key colums or similar operations which are changes the schema a lot). Migrations are good for development, but when you are updation dbase on UAT or in production, it is not possible to simple connection to Db from Visual Studio. You have to generate migration Sql script. It can be simply done via EF too. But the problem are static data inserted into Db in the Seed method. This method is called after applying last migration by VS.

A simple example of Seed method content:

  protected override void Seed(SampleContext context)
            //user are not filled automatically

            var roles = new List<Role>
                new Role{Id=1, Name="Admin", Description="System administration"},
                new Role{Id=2, Name="BasicUser", Description="User with low rights"},
                new Role{Id=3, Name="AdvanceUser", Description="User with advance rights"}

            roles.ForEach(e => context.Roles.Add(e));

            var countries = new List<Country>
                new Country{Id=1, Name="France"},
                new Country{Id=2, Name="Germany"},
                new Country{Id=3, Name="Austria"},
                new Country{Id=4, Name="Czech Republic"},
                new Country{Id=5, Name="Spain"},
                new Country{Id=6, Name="Norway"}
                //next dozens of countries

            countries.ForEach(e => context.Countries.Add(e));

But it is not possible to generate Sql inserts like scripts for transformation.
One solution can be to define data in seed method and next define Sql inserts in a
separate file. But this is an error prone solution when you forget to update one
of these sources. Next problem can be rewiting hundreds of lines (e.g. list of all
countries on the world).
For this purpose in our project, I have created a solution base on Xml and T4
template system. You simply define all static data in an Xml file. You can read
this file in Seed method and T4 template will generate Sql inserts from the
same Xml file too – sample diagram is depicted in followinf figure.

Process diagram

Sample Xml file with static data. In table element you define name of the table
(used by T4). includeId attribute is used to check if Id column should be set
into insert or not (because of autoincrement possibility). Last special thing is
postfix -n used to signalize the type of the column (to set ” around the value
or not).

<?xml version="1.0" encoding="utf-8" ?>
  <table includeId="true" name="Roles">
    <item id-n="1" name="Admin" description="System administration" />
    <item id-n="2" name="BasicUser" description="User with low rights" />
    <item id-n="3" name="AdvanceUser" description="User with advance rights" />
  <table includeId="false" name="Countries">
    <item name="France" />
    <item name="Germany" />
    <item name="Austria" />
    <item name="Czech Republic" />
    <!-- next dozens of countries -->

Source of the T4 template:

<#@ template debug="true" hostspecific="true" language="C#" #>
<#@ assembly name="System" #>
<#@ assembly name="System.Core" #>
<#@ assembly name="System.IO" #>
<#@ assembly name="System.Xml" #>
<#@ assembly name="System.Xml.Linq" #>
<#@ import namespace="System" #>
<#@ import namespace="System.Text" #>
<#@ import namespace="System.IO" #>
<#@ import namespace="System.Linq" #>
<#@ import namespace="System.Text" #>
<#@ import namespace="System.Xml.Linq" #>
<#@ import namespace="System.Collections.Generic" #>
<#@ output extension=".sql" #>

   <#       var path = Path.GetDirectoryName(Host.TemplateFile);
            path = Path.Combine(Path.GetDirectoryName(Host.TemplateFile),file);

            var doc = XDocument.Load(path);
            var root = doc.Element(dataElement);

            foreach (var table in root.Elements())
                foreach (var item in table.Elements())
                    var res = FormatInsertStatement(item, table.Attribute(nameAttribute).Value,
                    #><#= res #><#
                #><#= Environment.NewLine #><#


        const string file = "DbData.xml";
        const string nameAttribute = "name";
        const string includeIdAttribute = "includeId";
        const string idAttribute = "id";
        const string dataElement = "data";
        const string attributeNumberType = "-n";
        const string attributeTypeDelimiter = "-";
        string idNumberAttribute = string.Format("{0}{1}", idAttribute, attributeNumberType);

        private Tuple<string, string> FormatAttribute(XAttribute att, bool isLast)
            var name = att.Name.ToString();
            var value = att.Value;
            if (!name.EndsWith(attributeNumberType))
                value = string.Format("'{0}'", value);

                name = name.Remove(name.IndexOf(attributeTypeDelimiter));

            if (!isLast)
                value = value + ",";
                name = name + ",";
            return new Tuple<string, string>(name, value);

        private string FormatInsertStatement(XElement elem, string tableName, bool includeId)
            var attrs = elem.Attributes().ToArray();
            var attrStr = new StringBuilder(string.Format("INSERT INTO {0}(", tableName));
            var valStr = new StringBuilder("VALUES(");

            for (var i = 0; i < attrs.Length; ++i)
                var attr = attrs[i];
                var attrName = attr.Name.ToString().ToUpper();
                if (!includeId && (attrName.Equals(idAttribute.ToUpper()) || attrName.Equals(idNumberAttribute.ToUpper())))
                var res = FormatAttribute(attr, attrs.Length == i + 1);

            return string.Format("{0}) {1});{2}", attrStr.ToString(), valStr.ToString(), Environment.NewLine);


Result of the template applied on the sample Xml file:

INSER INTO Roles(id,name,description) VALUES(1,'Admin','System administration');
INSER INTO Roles(id,name,description) VALUES(2,'BasicUser','User with low rights');
INSER INTO Roles(id,name,description) VALUES(3,'AdvanceUser','User with advance rights');

INSER INTO Countries(name) VALUES('France');
INSER INTO Countries(name) VALUES('Germany');
INSER INTO Countries(name) VALUES('Austria');
INSER INTO Countries(name) VALUES('Czech Republic');

Finallz updated Seed method using Xml file with static data:

        private const string DataFile = "DbData.xml";
        private const string dataElement = "data";

        protected override void Seed(SampleContext context)
            var doc = XDocument.Load(DataFile);
            var root = doc.Element(dataElement);

            //user are not filled automatically

            var roles = from elem in GetTableData(root, "Roles").Elements()
                        select new Role
                            Id = int.Parse(elem.Attribute("id-n").Value),
                            Name = elem.Attribute("name").Value,
                            Description = elem.Attribute("description").Value

            roles.ToList().ForEach(e => context.Roles.Add(e));

            var countries = from elem in GetTableData(root, "Countries").Elements()
                        select new Country
                            Name = elem.Attribute("name").Value,

            roles.ToList().ForEach(e => context.Roles.Add(e));

        private XElement GetTableData(XElement root, string tableName)
            return root.Elements().Single(e => e.Attribute("name").Value.Equals(tableName));