Carl Hörberg on development

Automatic backup of Heroku database to S3

Tue, 17 May 2011 12:34:00 -0700

I didn't find any good solutions on how to automatically backup the Heroku database to S3, but with some inspiration from this blog post I cooked up a solution which takes advantage of the PG Backups and Cron addons on Heroku to backup to S3 daily, for free.

require 'aws/s3'
require 'heroku'
require 'heroku/command'
require 'heroku/command/auth'
require 'heroku/command/pgbackups'

task :cron do
  class Heroku::Auth
    def self.client
      Heroku::Client.new ENV['heroku_login'], ENV['heroku_passwd']
    end
  end

  Heroku::Command.run 'pgbackups:capture', ['--expire', '--app', ENV['heroku_appname']]
  url = capture_stdout do
    Heroku::Command.run 'pgbackups:url', ['--app', 'dmks']
  end

  AWS::S3::Base.establish_connection!(:access_key_id => ENV['s3_access_key'], :secret_access_key => ENV['s3_secret_key'])
  AWS::S3::S3Object.store "pgdb-#{Time.now.strftime('%y%m%d_%H%M')}.dump", open(url), ENV['s3_backup_bucket']
end

module Kernel
  def capture_stdout
    out = StringIO.new
    $stdout = out
    yield
    return out.string
  ensure
    $stdout = STDOUT
  end
end


            view raw
            Rakefile.rb
            This Gist brought to you by GitHub.
          

I hade to monkey patch Heroku::Auth so it doesn't ask for login/password from the prompt, and I had to add a 'capture_stdout' to Kernel to be able to get the url I then download and upload to S3.

Update: As Will Leinweber mentioned in the comments, Heroku are about to offer this functionallity, but to date this feature is still in private beta.

How to list enum values in DataMapper

Tue, 17 May 2011 02:53:00 -0700

The Ruby ORM DataMapper as a Enum property which is really handy, but you often want to list the available options in for example a form. Unfortunately it's not well documented, but this gist shows you how:

class Entity
  include DataMapper::Resource
  property :enum_property, Enum[:a, :b, :c]
end

Entity.enum_property.options[:flags] #=> [:a, :b, :c]
# or
Entity.enum_property.flag_option.values #=> [:a, :b, :c]


            view raw
            gistfile1.rb
            This Gist brought to you by GitHub.
          

Why Ruby is so much better than C#

Tue, 03 May 2011 23:17:00 -0700

I've been thinking a lot about why my development team is so much more productive since we switched to Ruby/Sinatra from C#/Asp.Net MVC. Do we get stuck less often? No, learning new libs always takes times and have some quirkies. Are we better Ruby programmers than C# progammers? Most unlikely, at least we got a lot of more C# experince. So what is it? I think Karl Seguin can be on to something:

I don't think C#/Java developers have any concept of how much energy/productivity they spend on decoupling. Worse, I think many C#/Java developers don't understand that DI, interface-based programming, and fear of statics, is result of language limitations. -- Karl Seguin

In C# you have to have so much infrastructure, and you tend to work a lot with it. You put a lot of effort to not use static, not coupling classes etc. because then you are in a lot of trouble when it comes to testing. In Ruby it's so much easier to just get things done. No Visual Studio bloat, adding a ton of libs, wiring up DI, separting you domain models from you ORM etc. And because of the simplicity of the language constructors you don't miss VisualStudio, IntelliSense or ReSharper. The speed you get with Vim when using Ruby is staggering.

New blog

Tue, 03 May 2011 12:29:00 -0700

I justed moved my blog from Posterous, rolled my own blog engine and put it on Heroku.

The architecture is really simple, it's a plain Sinatra app which just reads the posts from a static YAML file. Of course it want scale if you blog like Ayende Rahien but for me it will do until later. Comments are provided through Disqus. The posts is formatted in Markdown. There are several reasons why I like having my posts under source control:

No need for database backup
I can use Vim to write my posts
The site is read-only, so no security mesaures is needed.
No editor or domain models is needed in the app.

You can see the code over at GitHub.

How to unit test RabbitMQ in C#

Sun, 02 Jan 2011 13:01:00 -0800

RabbitMQ is a very good open source message queue server, but unfortunately it's DotNet API isn't very "dotnetish". Well, finally I figured out a way to unit test classes which relied upon IModel.

[TestFixture]
public class RabbitMQTest
{
    [Test]
    public void CanMockMQ()
    {
        // Arrange
        var channelMock = new Mock<IModel>();
        channelMock
            .Setup(m => m.BasicConsume(It.IsAny<string>(), It.IsAny<bool>(), It.IsAny<IDictionary>(), It.IsAny<IBasicConsumer>()))
            .Callback<string, bool, object, IBasicConsumer>((queue, noAck, props, consumer) => consumer.HandleBasicDeliver("", 1, false, "exchange", "route", null, new byte[0]));
        var server = new MQUser(channelMock.Object);

        // Act
        server.Abort();
        server.DoStuff();

        // Assert
        channelMock.Verify(m => m.BasicAck(1, false));
    }
}

            view raw
            Test.cs
            This Gist brought to you by GitHub.
          

public class MQUser
{
    readonly IModel channel;

    public MQUser(IModel channel)
    {
        this.channel = channel;
    }

    public void DoStuff()
    {
        var queueName = channel.QueueDeclare();
        var subscription = new Subscription(channel, queueName);
        do
        {
            var msg = subscription.Next();
            // Do stuff with the message
            subscription.Ack(msg);
        } while (!aborted);
    }

    bool aborted;
    public void Abort()
    {
        aborted = true;
    }
}


            view raw
            TestSubject.cs
            This Gist brought to you by GitHub.
          

It uses Moq which let us mock the IModel interface and intercepts all calls to IModel#BasicConsume, which all subscribers must call to subscribe for messages, and then imeditly call IModel#HandleBasicDeliver which propagetes that a messages has arrived to the subscriber. Then the test continues by starting the class under test which contains a simple subscriber in a abortable loop. We will abort it before starting the loop (but becasue it's a do-while loop, it will try to fetch one message before returning). At last we verify that the message has been acknowledged.

iPad app tips

Sun, 12 Dec 2010 14:58:00 -0800

Quite a few friends are equipping them self with iPad's now that they finally are available in Sweden, and while I've had one for 6-7 months by now I though it could be good idea to list my most frequently used apps.

Spotify (iPhone), with iOS 4.2 (multitasking) this app is a must.
Wikipanion, wikipedia client.
IMDb, very good high quality IMDb client.
Heritage, photos from all world heritages sites.
Skype, the iPad's built in mic and speakers are very high quality.
Instapaper, when ever i encounter long but interresting articles I mark them as "Read leader" in Chrome, and then use this app to read it because of the much more pleasant reading experience on the pad.
Angry Birds, doesn't need any introduction..
Cut the Rope, a game in the same spirit as Angry Birds. Easy and entertaining.
Monkey Island 2 Special Edition, better graphics, voice dialogs, same good story.
Dreams, find differences between two pictures, child game which also adults enjoys.
Sudoku Tablet, for the Sudoku lovers.
iBooks, the PDF and eBook reader.
Kindle, get access to the full Amazon book supply.
River of News, best RSS/Google Reader app I've found.
Twitter, best twitter client, in all categories. Actually makes Twitter usable.
Line Art, meaningless but relaxing and beautiful app where you control a myriad of lines/dots with your fingers.
LIFE, breathtaking photos from the american Life magazine. I can browse it for hours.
Dropbox, file backup/sharing.

Swedish

DN+, a app which actually embrace the iPad (unlike for example SVD which merely is a glorified PDF reader).
ÖGT (iPhone), unexpectedly good local traffic app (for Östergötland).
Krysset, crossword game from Bonnier.

Apps which ought to be good but which I've haven't used a lot.

TED, great, short, insightful videos.
VLC, play videos of formats the iPad normally don't support, like xvid (you still need iTunes to transfer the videos, which is a show blocker for me).
Trapster, warnings for speed cameras/police checkpoints, crowd sourced. Hard to determine the reliability.
RDP Lite, Windows Remote Desktop client for those who have to administer windows servers.
Flipboard, very hype. Shows of Twitter and other sources as a traditional newspaper.
Eyewithness, beautiful and engaging photos from the Guardian.
Grand Theft Auto, CTW Lite. Quite entertaining (the one time I played it).
IM+, the iPad don't lend it self for a lot of keyboard typing, but when you have to chat IM+ is my choice.
The Economist, the free edition only allows you the read Editor picks, but the app is very high quality and so is of course the content.
eBokus (Swedish), swedish books.
Trafikinfo (Swedish), traffic information, great to be aware of accidents or road works before you're in the queue.

Twitter + Beanstalk + EventMachine + WebSocket

Wed, 13 Oct 2010 11:05:00 -0700

Justed forked a Twitter AMQP WebSocket example project by Ruben Fonseca but replaced the AMQP/RabbitMQ with Beanstalk instead. I've been playing with this fast and simple message queue called Beanstalk for a couple of days and now wanted to see if I could do some WebSocket work with it as well. You can find an introduction to Beanstalk here and some comparison to other MQs here.

The EventMachine, which is a highly scalable, event driven i/o library for ruby (and c++ and java), was a very nice surprise and was very easy to work with. Both the Twitter Streaming library, the Beanstalk client and the WebSocket server uses EventMachine so the whole example it fully asynchronous and probably quite scalable.

Check it out: http://github.com/carlhoerberg/twitter-beanstalk-websocket-example

How to override ActiveRecord primary key and foreign key conventions

Thu, 07 Oct 2010 04:40:00 -0700

While evaluating the Microsoft SQL Server adapter for ActiveRecord and IronRuby i wanted the ability to override some activerecord conventions, class wide, so I didn't have to override it in every class and every association. Specifically I wanted the foreign keys to be named "classNameId" and not "className_id" and the primary keys "Id" instead of lower case "id", this is how you do it:

module ActiveSupport
  module Inflector
    alias :old_foreign_key :foreign_key
    def foreign_key(class_name, separate_class_name_and_id_with_underscore = true)
      old_foreign_key(class_name, false)
    end
    def tableize(class_name)
      class_name
    end
  end
end

module ActiveRecord
  class Base
    class << self
      def get_primary_key(base_name) #:nodoc:
        "Id"
      end
    end
  end
end

            view raw
            gistfile1.rb
            This Gist brought to you by GitHub.
          

My development stack

Sat, 11 Sep 2010 10:45:00 -0700

This is the development stack I employ, and why I chosed it:

ASP.NET MVC

Being mostly a C# programmer (altough Ruby is growing on me) ASP.NET MVC is a pretty natural selection, being the most popular MVC framework for .NET. I don't think _not_ using a MVC framework for web development is very serious today. A MVC framework got so many benefits compared to for example WebForms I can't even begin, so i let Karl Seguin explain MVC vs. WebForms instead.

Autofac

In a static language like C# you can't live without a dependency injection framework, i choose Autofac a long time ago and have stuck with it, i think most IoC frameworks are on par in the .NET world, but it seems to me that Autofac most often have the edge, coming up with new compelling features first.

nUnit

Proven unit testing framework. Tried xUnit for a while but the poor tool support (in ReSharper and TeamCity) made me go back to nUnit, it just works.

Moq

There have been a lot of discussions on wether to use a mocking framework or not, but I like to write as lite code as possible when testing so a mocking is a perfect fit for me.

SpecFlow

The Ruby community came up with Cucumber in which you write specifications in plain english (or swedish), and have become very successful. The idea has now been ported to .NET. Our customers love it!

WatiN

For integration testing we use WatiN which automates Internet Explorer (and Firefox).

AutoMapper

Mapping code is boring and error prone, automapper lets us do the entity->viewmodel mapping by ease.

Elmah

Web error logging module, drop it in and all unhandled exceptions are mailed to us.

log4net

Proven logging framework for .NET, we use it in all apps for error and trace logging. Although, it's pretty ugly to configure through code.

PdfSharp

PDF creator, a lot better API than iTextSharp, but don't got all the features of it.

NPOI

Excel file creator, ported for java's POI, which the API suffers from, but got a lot of features (except rendering charts dynamically, which is a bummer.

T4MVC

Instead of referring to views and actions in code with "magic string" T4MVC creates strongly typed helpers for you.

Linq to SQL

A simple ORM which don't got the impressive feature set or the performance of nHibernate, but do got a supreme LINQ engine. Read my post on how to get the most out of Linq to SQL.

MigratorDotNet

Don't do RMDBS with out a migration tool. It will ease a lot of the pain a static schema forces upon you. MigratorDotNet isn't maintained anymore but some alternatives are NMigrations and Fluent Migrator.

jQuery (and jQuery UI)

For all our javascript needs..

YUI Grids

CSS framework, but will probably use 960 Grid System next time.

TeamCity

Continuous integration has always been a natural part of your workflow. TeamCity is very feature rich build management and CI server. It listens at our source code repository for changes, when some one commits anything it builds it, run all unit and integration tests and then deploys it to the staging/production servers, all with out us moving a finger.

GitHub

Previously we used SVN with VisualSVN and hosted it ourself, but we've now employing Git, mostly because it's very good branch capabilities. Although a Git server is some what cumbersome to set up on a windows machine, so just signed up for GitHub, less hassel.

Cygwin

Unfortunately the Git support for windows is quite bad, so we use Cygwin (which "emulates" the linux api) to get the power of all those linux commands and applications, including Git right at our windows desktops.

Cygwin git svn unable to remap?

Wed, 08 Sep 2010 01:21:00 -0700

Getting errors like the one below in Cygwin?

  2 [main] perl 8496 C:\cygwin\bin\perl.exe: \***| fatal error - unable to remap C:\cygwin\bin\cygsvn\_fs\_base-1-0.dll to same address as parent: 0xB00000 != 0xC10000   
  2 [main] perl 11192 fork: child 8496 - died waiting for dll loading, errno 11   
  2 [main] perl 8496 C:\cygwin\bin\perl.exe: \***| fatal error - unable to remap C:\cygwin\bin\cygsvn\_fs\_base-1-0.dll to same address as parent: 0xB00000 != 0xC10000   
  2 [main] perl 11192 fork: child 8496 - died waiting for dll loading, errno 11

Solution: Temporally disable your anti-virus software.

Efficient workflow with Linq to Sql using Migrator, T4 and some Xslt

Wed, 26 May 2010 03:44:00 -0700

Linq to SQL is still my favorite datalayer, why? Excellent Linq capabilities and the fact that I only have to change my database once, and my code/entities will be immediately reflected. How?

First, download MigratorDotNet, it's a utility for database migrations, which versions your database changes in code: http://code.google.com/p/migratordotnet/wiki/GettingStarted. Create a new library in your solution and start writing migrations.

Don't let VS generate your datacontext and entities, L2ST4 does it a lot better, essentially it's T4 templates which allows you to change exactly the way your datacontext and entities are generated.

Now copy and add

@echo off
setlocal
set path="%CommonProgramFiles(x86)%\\microsoft shared\\TextTemplating\\10.0";"%ProgramFiles(x86)%\\Microsoft SDKs\\Windows\\v7.0A\\Bin\\NETFX 4.0 Tools";%path%

set connStr="Data Source=.\\SQLEXPRESS;Initial Catalog=MyDB;Integrated Security=true"

REM Download MigratorDotNet from http://code.google.com/p/migratordotnet/
echo Applying migrations
..\\Libs\\Migrator\\Migrator.Console SqlServer %connStr% "..\\MyProject.Migrations\\Bin\\Debug\\MyProject.Migrations.dll" -trace

REM If you only have Windows SDK 6 then change the path in top to "%programfiles%\\Microsoft SDKs\\Windows\\v6.0A\\Bin"
echo Regenerating DBML
sqlmetal /conn:%connStr% /dbml:MyDB.dbml /namespace:MyProject.Data.Entities /context:MyDBDataContext /functions 

REM Download msxsl.exe from http://www.microsoft.com/downloads/details.aspx?familyid=2fb55371-c94e-4373-b0e9-db4816552e41&displaylang=en
echo Applying XSLT to DBML and copying to destination
msxsl MyDB.dbml MyDB.xslt -o ..\\MyProject.Data\\Entities\\MyDB.dbml

REM Download L2ST4 from http://l2st4.codeplex.com/
REM If having trouble with TextTransform, change path in top to to "%CommonProgramFiles(x86)%\\microsoft shared\\TextTemplating\\1.2"
echo Transforming L2ST4 template
TextTransform ..\\MyProject.Data\\Entities\\MyDB.tt

echo Done
endlocal


            view raw
            MigrateAndGenerateContext.cmd
            This Gist brought to you by GitHub.
          

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet version="1.0"
	xmlns:dbml="http://schemas.microsoft.com/linqtosql/dbml/2007"
	xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
	xmlns:msxsl="urn:schemas-microsoft-com:xslt"
        xmlns:user="urn:script-items"
	exclude-result-prefixes="msxsl dbml user">
  <xsl:output method="xml" indent="yes"/>

  <!-- By default copy all nodes an attributes -->
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
    </xsl:copy>
  </xsl:template>

  <!-- You probably don't want an Entity for the SchemaInfo table which Migrator will add
       This is how you hide a table 
  -->
  <xsl:template match="//dbml:Table[@Name='dbo.SchemaInfo']"/>

  <!-- Replace a column with a Enum.
       The column in the MyTable table is called TypeEnum. 
       The Enum defined in my project is called "MyEnum".
       I replace the name "TypeEnum" with "Type"
  -->
  <xsl:template match="//dbml:Type[@Name = 'MyTable']/dbml:Column[@Name = 'TypeEnum']">
    <xsl:copy>
      <xsl:apply-templates select="@* | node()" />
      <xsl:attribute name="Type">MyEnum</xsl:attribute>
      <xsl:attribute name="Member">Type</xsl:attribute>
    </xsl:copy>
  </xsl:template>
  
  <!-- SqlMetal has a bug 
       (http://connect.microsoft.com/VisualStudio/feedback/details/486598/sqlmetal-generates-incorrectly-capitalized-identifiers)
       which this code will resolve, only needed if you have non ascii characters in you table or column names.
       But then you need a better MSXSL.exe: http://www.codeproject.com/KB/XML/msxslplus.aspx
  -->
  <!--
  <xsl:template match="@Name">    
    <xsl:attribute name="Name">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@Member">
    <xsl:attribute name="Member">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@Type">
    <xsl:attribute name="Type">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@ThisKey">
    <xsl:attribute name="ThisKey">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@OtherKey">
    <xsl:attribute name="OtherKey">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>

  <msxsl:script language="C#" implements-prefix="user">
    <![CDATA[
    public string FixCasing(string str)
    {
      char[] chars = str.ToCharArray();
      
      for (int i = 0; i < chars.Length - 1; i++)
        // Add your troubling characters
        if ((chars[i] == 'å' || chars[i] == 'ä' || chars[i] == 'ö' || chars[i] == 'Å' || chars[i] == 'Ä' || chars[i] == 'Ö') && char.IsLetter(chars[i + 1]))
          chars[i + 1] = char.ToLower(chars[i + 1]);

      return new string(chars);
    }
    ]]>
  </msxsl:script>
  -->
</xsl:stylesheet>


            view raw
            MyDB.xslt
            This Gist brought to you by GitHub.
          

#file_migrate_and_generate_context.cmd">MigrateAndGenerateContext.cmd and

@echo off
setlocal
set path="%CommonProgramFiles(x86)%\\microsoft shared\\TextTemplating\\10.0";"%ProgramFiles(x86)%\\Microsoft SDKs\\Windows\\v7.0A\\Bin\\NETFX 4.0 Tools";%path%

set connStr="Data Source=.\\SQLEXPRESS;Initial Catalog=MyDB;Integrated Security=true"

REM Download MigratorDotNet from http://code.google.com/p/migratordotnet/
echo Applying migrations
..\\Libs\\Migrator\\Migrator.Console SqlServer %connStr% "..\\MyProject.Migrations\\Bin\\Debug\\MyProject.Migrations.dll" -trace

REM If you only have Windows SDK 6 then change the path in top to "%programfiles%\\Microsoft SDKs\\Windows\\v6.0A\\Bin"
echo Regenerating DBML
sqlmetal /conn:%connStr% /dbml:MyDB.dbml /namespace:MyProject.Data.Entities /context:MyDBDataContext /functions 

REM Download msxsl.exe from http://www.microsoft.com/downloads/details.aspx?familyid=2fb55371-c94e-4373-b0e9-db4816552e41&displaylang=en
echo Applying XSLT to DBML and copying to destination
msxsl MyDB.dbml MyDB.xslt -o ..\\MyProject.Data\\Entities\\MyDB.dbml

REM Download L2ST4 from http://l2st4.codeplex.com/
REM If having trouble with TextTransform, change path in top to to "%CommonProgramFiles(x86)%\\microsoft shared\\TextTemplating\\1.2"
echo Transforming L2ST4 template
TextTransform ..\\MyProject.Data\\Entities\\MyDB.tt

echo Done
endlocal


            view raw
            MigrateAndGenerateContext.cmd
            This Gist brought to you by GitHub.
          

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet version="1.0"
	xmlns:dbml="http://schemas.microsoft.com/linqtosql/dbml/2007"
	xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
	xmlns:msxsl="urn:schemas-microsoft-com:xslt"
        xmlns:user="urn:script-items"
	exclude-result-prefixes="msxsl dbml user">
  <xsl:output method="xml" indent="yes"/>

  <!-- By default copy all nodes an attributes -->
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
    </xsl:copy>
  </xsl:template>

  <!-- You probably don't want an Entity for the SchemaInfo table which Migrator will add
       This is how you hide a table 
  -->
  <xsl:template match="//dbml:Table[@Name='dbo.SchemaInfo']"/>

  <!-- Replace a column with a Enum.
       The column in the MyTable table is called TypeEnum. 
       The Enum defined in my project is called "MyEnum".
       I replace the name "TypeEnum" with "Type"
  -->
  <xsl:template match="//dbml:Type[@Name = 'MyTable']/dbml:Column[@Name = 'TypeEnum']">
    <xsl:copy>
      <xsl:apply-templates select="@* | node()" />
      <xsl:attribute name="Type">MyEnum</xsl:attribute>
      <xsl:attribute name="Member">Type</xsl:attribute>
    </xsl:copy>
  </xsl:template>
  
  <!-- SqlMetal has a bug 
       (http://connect.microsoft.com/VisualStudio/feedback/details/486598/sqlmetal-generates-incorrectly-capitalized-identifiers)
       which this code will resolve, only needed if you have non ascii characters in you table or column names.
       But then you need a better MSXSL.exe: http://www.codeproject.com/KB/XML/msxslplus.aspx
  -->
  <!--
  <xsl:template match="@Name">    
    <xsl:attribute name="Name">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@Member">
    <xsl:attribute name="Member">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@Type">
    <xsl:attribute name="Type">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@ThisKey">
    <xsl:attribute name="ThisKey">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@OtherKey">
    <xsl:attribute name="OtherKey">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>

  <msxsl:script language="C#" implements-prefix="user">
    <![CDATA[
    public string FixCasing(string str)
    {
      char[] chars = str.ToCharArray();
      
      for (int i = 0; i < chars.Length - 1; i++)
        // Add your troubling characters
        if ((chars[i] == 'å' || chars[i] == 'ä' || chars[i] == 'ö' || chars[i] == 'Å' || chars[i] == 'Ä' || chars[i] == 'Ö') && char.IsLetter(chars[i + 1]))
          chars[i + 1] = char.ToLower(chars[i + 1]);

      return new string(chars);
    }
    ]]>
  </msxsl:script>
  -->
</xsl:stylesheet>


            view raw
            MyDB.xslt
            This Gist brought to you by GitHub.
          

#file_my_db.xslt">MyDB.xslt from the code samples below to your solution and run the cmd (here is a solution how to run .cmd files from within VS). What it will do is to apply your migrations, eg. Add tables, rename columns etc., then generate a new .dbml file using SqlMetal, then the XSLT will transform the .dbml (which simply is a XML file) and it will apply your changes like replacing some columns with your Enums or hiding some tables you don't want classes for. You need a Xslt transformer for that step to work, you can download msxsl.exe from Microsoft or a version which can do inline scripts from here, which you need if you want to take use of the SqlMetal bugfix i've supplied in the xslt. At last the script will automatically run the L2ST4 template and generate the new datacontext and entities, all without manual/error prone interaction.

Done! You've changed your database schema and regenerated the correct datacontext and entities with one code change and one click!

Now all the changes you've done is applied at your developer machine, now add the Migrator.Console.exe part to your deploy script and then your live database will be updated automatically. No more tedious executing .sql files manually or keeping track of which .sql file is applied or not!

@echo off
setlocal
set path="%CommonProgramFiles(x86)%\\microsoft shared\\TextTemplating\\10.0";"%ProgramFiles(x86)%\\Microsoft SDKs\\Windows\\v7.0A\\Bin\\NETFX 4.0 Tools";%path%

set connStr="Data Source=.\\SQLEXPRESS;Initial Catalog=MyDB;Integrated Security=true"

REM Download MigratorDotNet from http://code.google.com/p/migratordotnet/
echo Applying migrations
..\\Libs\\Migrator\\Migrator.Console SqlServer %connStr% "..\\MyProject.Migrations\\Bin\\Debug\\MyProject.Migrations.dll" -trace

REM If you only have Windows SDK 6 then change the path in top to "%programfiles%\\Microsoft SDKs\\Windows\\v6.0A\\Bin"
echo Regenerating DBML
sqlmetal /conn:%connStr% /dbml:MyDB.dbml /namespace:MyProject.Data.Entities /context:MyDBDataContext /functions 

REM Download msxsl.exe from http://www.microsoft.com/downloads/details.aspx?familyid=2fb55371-c94e-4373-b0e9-db4816552e41&displaylang=en
echo Applying XSLT to DBML and copying to destination
msxsl MyDB.dbml MyDB.xslt -o ..\\MyProject.Data\\Entities\\MyDB.dbml

REM Download L2ST4 from http://l2st4.codeplex.com/
REM If having trouble with TextTransform, change path in top to to "%CommonProgramFiles(x86)%\\microsoft shared\\TextTemplating\\1.2"
echo Transforming L2ST4 template
TextTransform ..\\MyProject.Data\\Entities\\MyDB.tt

echo Done
endlocal


            view raw
            MigrateAndGenerateContext.cmd
            This Gist brought to you by GitHub.
          

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet version="1.0"
	xmlns:dbml="http://schemas.microsoft.com/linqtosql/dbml/2007"
	xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
	xmlns:msxsl="urn:schemas-microsoft-com:xslt"
        xmlns:user="urn:script-items"
	exclude-result-prefixes="msxsl dbml user">
  <xsl:output method="xml" indent="yes"/>

  <!-- By default copy all nodes an attributes -->
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
    </xsl:copy>
  </xsl:template>

  <!-- You probably don't want an Entity for the SchemaInfo table which Migrator will add
       This is how you hide a table 
  -->
  <xsl:template match="//dbml:Table[@Name='dbo.SchemaInfo']"/>

  <!-- Replace a column with a Enum.
       The column in the MyTable table is called TypeEnum. 
       The Enum defined in my project is called "MyEnum".
       I replace the name "TypeEnum" with "Type"
  -->
  <xsl:template match="//dbml:Type[@Name = 'MyTable']/dbml:Column[@Name = 'TypeEnum']">
    <xsl:copy>
      <xsl:apply-templates select="@* | node()" />
      <xsl:attribute name="Type">MyEnum</xsl:attribute>
      <xsl:attribute name="Member">Type</xsl:attribute>
    </xsl:copy>
  </xsl:template>
  
  <!-- SqlMetal has a bug 
       (http://connect.microsoft.com/VisualStudio/feedback/details/486598/sqlmetal-generates-incorrectly-capitalized-identifiers)
       which this code will resolve, only needed if you have non ascii characters in you table or column names.
       But then you need a better MSXSL.exe: http://www.codeproject.com/KB/XML/msxslplus.aspx
  -->
  <!--
  <xsl:template match="@Name">    
    <xsl:attribute name="Name">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@Member">
    <xsl:attribute name="Member">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@Type">
    <xsl:attribute name="Type">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@ThisKey">
    <xsl:attribute name="ThisKey">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>
  <xsl:template match="@OtherKey">
    <xsl:attribute name="OtherKey">
      <xsl:value-of select="user:FixCasing(.)"/>
    </xsl:attribute>
  </xsl:template>

  <msxsl:script language="C#" implements-prefix="user">
    <![CDATA[
    public string FixCasing(string str)
    {
      char[] chars = str.ToCharArray();
      
      for (int i = 0; i < chars.Length - 1; i++)
        // Add your troubling characters
        if ((chars[i] == 'å' || chars[i] == 'ä' || chars[i] == 'ö' || chars[i] == 'Å' || chars[i] == 'Ä' || chars[i] == 'Ö') && char.IsLetter(chars[i + 1]))
          chars[i + 1] = char.ToLower(chars[i + 1]);

      return new string(chars);
    }
    ]]>
  </msxsl:script>
  -->
</xsl:stylesheet>


            view raw
            MyDB.xslt
            This Gist brought to you by GitHub.
          

Asp.Net MVC and the Microsoft Chart Controls in .NET 4.0

Sun, 25 Apr 2010 04:53:00 -0700

I'm a big fan of Microsoft Chart Controls. In .NET 4 the Chart Controls has become a part of the framework so there is not a separate download as to .NET 3.5. In this post I will show how to integrate them with Asp.Net MVC on .NET 4 because there are some changes from .NET 3.5. Also the most examples on the web suggests that you should output the resulting image directly from the Controller, I will show you how to take use of the full capabilities of Chart Controls and let it render it's own html, including . Last in the post are all the code examples.

In .NET 4.0 the Chart Controls are integrated and with the reduced web.config you don't have to add that much stuff, but there is some, most notably the ChartImg.axd handler (see below).

A new thing quirk in .NET 4.0 is that the ChartImg.axd always will append to your current url. Therefor we need to ignore all urls ending with ChartImg.axd in our routetable. This ignore rule has to be applied before any Area registration, else wise it won't be applied in your Areas.

The core thing in this solution is that you preserve the Chart object as long as possible, and then let it render it self. That gives us the benefits of all extra html the Chart Control can generate (like imagemap) and image caching.

<%@ Control Language="C#" Inherits="System.Web.Mvc.ViewUserControl<System.Web.UI.DataVisualization.Charting.Chart>" %>
<%
    Model.Page = this.Page;
    var writer = new HtmlTextWriter(Page.Response.Output);
    Model.RenderControl(writer);
%>

            view raw
            Chart.ascx
            This Gist brought to you by GitHub.
          

public class ChartController: Controller
{
  public ActionResult MyChart()
  {
    var chart = new Chart();
    // Configure chart
    return View(chart);
  }
}

            view raw
            ChartController.cs
            This Gist brought to you by GitHub.
          

protected void Application_Start()
{
  // It's importat that this ignore of any url ending with ChartImg.axd is inserted before any Area route registrations.
  // If you won't use Charts in any Area you can put it in your RegisterRoutes() method. 
  RouteTable.Routes.Ignore("{*pathInfo}", new { pathInfo = @"^.*(ChartImg.axd)$" });
  AreaRegistration.RegisterAllAreas();
  RegisterRoutes(RouteTable.Routes);

  // The rest.. 
}


            view raw
            global.asax.cs
            This Gist brought to you by GitHub.
          

<% Html.RenderPartial("Chart", Model); %>

            view raw
            MyChart.aspx
            This Gist brought to you by GitHub.
          

<?xml version="1.0"?>
<configuration>

 <!-- Your stuff -->

 <appSettings>   
   <add key="ChartImageHandler" value="storage=file;timeout=20;" />
 </appSettings>

 <!-- Your stuff -->

 <system.web>
    <compilation debug="true" targetFramework="4.0">
      <assemblies>
       <!-- Other assemblies -->
        <add assembly="System.Web.DataVisualization, Version=4.0.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35"/>
      </assemblies>
    </compilation>

    <!-- Your stuff -->

    <httpHandlers>
      <!-- Other handlers -->
      <add path="ChartImg.axd" verb="GET,HEAD" type="System.Web.UI.DataVisualization.Charting.ChartHttpHandler, System.Web.DataVisualization, Version=4.0.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" validate="false"/>
    </httpHandlers>
 </system.web>

 <!-- Your stuff -->

 <system.webServer>
  <handlers>
   <!-- Your handlers -->
   <add name="ChartImageHandler" preCondition="integratedMode" path="ChartImg.axd" verb="GET,HEAD" type="System.Web.UI.DataVisualization.Charting.ChartHttpHandler, System.Web.DataVisualization, Version=4.0.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35"/>
  </handlers>
 </system.webServer>

 <!-- Your stuff -->

</configuration>

            view raw
            web.config.xml
            This Gist brought to you by GitHub.
          

How to use JQuery UI Autocomplete with Asp.Net MVC

Tue, 20 Apr 2010 11:51:00 -0700

Microsoft has swifted away from Ajax Library and they are promoting jQuery as the primary way of develop client side ajax applications. This is good news, Microsoft is even contributing resources to the jQuery core team.

A couple of weeks ago i showed how you could use Ajax Library Autocomplete component with Asp.Net MVC. This time I will instead show you how to make the new Autocomplete component in jQuery UI 1.8 play well with Asp.Net MVC, and it's a lot easier than with Ajax Library!

First, in your view, include jQuery 1.4 and jquery UI 1.8 (here i use Google's javascript CDN), then add a text field to your body, then some javascript to hook it up:

<html>
<head>
  <link href="http://ajax.googleapis.com/ajax/libs/jqueryui/1.8/themes/cupertino/jquery-ui.css" type="text/css" rel="Stylesheet" />
  <script src="http://ajax.googleapis.com/ajax/libs/jquery/1.4/jquery.min.js" type="text/javascript"></script>
  <script src="http://ajax.googleapis.com/ajax/libs/jqueryui/1.8/jquery-ui.min.js" type="text/javascript"></script>
</head>
<body>
<input type="text" name="completeMe" id="completeMe" />
<script type="text/javascript">
$(function() {
            $("#completeMe").autocomplete({
                source: '/Autocomplete/Users',
                minLength: 1,
                select: function(event, ui) {
                    // Do something with  "ui.item.Id" or "ui.item.Name" or any of the other properties you selected to return from the action 
                }
            });
});
</script>
</body>
</html>

            view raw
            autocomplete.html
            This Gist brought to you by GitHub.
          

    public class AutocompleteController : Controller
    {
        private readonly IRepository db;
        public AutocompleteController(IRepository db)
        {
            this.db = db;
        }

        public JsonResult Users(string term)
        {
            var data = db.Query<User>()
                .Where(r => r.Name.StartsWith(term))
                .Select(r => new { r.Id, value = r.Name })
                .Take(10)
                .ToArray();
            return Json(data, JsonRequestBehavior.AllowGet);
        }


            view raw
            AutocompleteController.cs
            This Gist brought to you by GitHub.
          

I've included the select event, which will fire when the users has selected an item, in which you can load more stuff etc. Be sure to check out the API for autocomplete for more events and properties.

If you look at the Autocomplete controller you see that I query my User table where the Users name starts with the incoming term. We select only the id and the name from the user (you can select more stuff if you want to make it available in your jQuery events). I only take 10 items, make an array of them which we return as Json.

Important to note is that we have to add "JsonRequestBehavior.AllowGet" argument to the Json() function because jQuery UI Autocomplete makes a Get request and Asp.Net MVC by default doesn't allow this because of the security concerns.