This walkthrough will take you through the steps to get a Merb app up and running with a protected resource. It assumes you are running Merb 0.9.9 or greater with the Merb Stack (merb-core, merb-more, and datamapper).

Lets start at the beginning and create an app
$ merb-gen app authentication_app
$ cd authentication_app

By default merb-gen app will create a user model for you in app/models/user.rb. The model is there, but we still need to create the table in the database and add a user to authenticate with. Lets do that now.
$ rake db:auto_migrate

To create our first user we can drop into Merb’s irb console mode with merb -i. Like Rails’ script/console, merb -i gives you access to your models and other classes.

$ merb -i
u = User.new
u.login = 'joe'
u.password = u.password_confirmation = 'password'
u.save
exit

In order to test authentication we need something to protect. Any controller would work, but lets create a resource.
$ merb-gen resource secret

We will need to add a route to the resource also
# router.rb
Merb::Router.prepare do
resources :secrets
...
end

To protect a controller with authentication we can use the before filter ensure_authenticated
# app/controllers/secrets.rb
class Secrets < Application
before :ensure_authenticated
...
end

Everything should now be setup in our app. Lets give it a run through.

First you need to start you merb server. Make sure you are in your merb app’s root directory.
$ merb

Now you can test access to your resource is denied before authentication.
http://localhost:4000/secrets

Login in and try to access the recourse again.
http://localhost:4000/login
http://localhost:4000/secrets

You should now be able to access the resource. Awesome.

When I was somewhere between 8-10 years old my uncle gave my family our first computer, a Coleco Adam. Along with the computer we were given the now classic book BASIC Computer Games. No one in my family new anything about compilers, programming, or computers in general, but we were excited about the idea of creating something out of nothing so decided to give one of these programs a go.

We decided on Bunny and began to slowly peck out the source code into the text editor. We saved the file, exited the editor, and then stared blankly at the screen as we realized we didn’t know what to do next. Interpreters and compilers were obviously never discussed and after our futile attempts of running our program failed, I decided to take a different approach.

What was I trying to accomplish? What were the outcomes I needed? The answer was simple- an ASCII image of a bunny. So that’s what I created, and ASCII image of a bunny. For the rest of the afternoon I typed out the bunny by hand and by using brute force I eventually accomplished the goal.

This retrospective reminds me that when things “just aren’t working”, sometimes I need to take a step back and reevaluate my methods. I may have an Occam’s razor type revelation that could satisfy requirements without significant compromises.

Keep it old school and tell the world your first programming experience in a comment below.

Merb Overheard follows the planet-type aggregation sites similar to Planet Ruby on Rails and Planet PHP. It follows a list of relevant blogs and aggregates those articles into one place.

When I first decided to build Merb Overheard I was originally planning on writing it in Merb of course, but I starting to realize all the battles I was going to have ahead of me. Caching and pagination of no database objects were not something I really wanted to deal with, so I decided to try something different. Instead of trying to shove the aggregator into a framework, why not just build out the pages statically?

Merb Overhead is just a plain old ruby script that gets ran every 30 minutes from a cron task. Each time it is ran, all the pages are recreated using Erubius. This script, weighing in at a whopping 8 lines of code, implements two classes in a library I wrote called Pluto. The class for the feed aggregator is only 30 lines. The class to generate the view is only 45 lines.

The code is released on Github so go nuts creating collecting all your Furby Fan Club sites or whatever else strikes your fancy.

./script/plugin install git://github.com/technoweenie/restful-authentication.git -r modular

This will install the modular branch of the Restful Authentication plugin. Pretty cool, huh?

At the time of writing this Rails 2.2 has not been released. If you are using Rails 2.1.0 or 2.1.1 you will need to upgrade to Edge Rails for this to work. Otherwise you will have to wait for 2.2 to be released.

First off let me say this about Feather. It is an amazing application. I have contributed a bit to its development and can vouch on the quality of the Product. The creators of Feather, Michael Leung and El Draper, have done a great job putting together a wonderful blogging tool that is both simple and extendable. The team has had the difficult task of trying to keep of with the rapidly mutating Merb and DataMapper code bases and have managed to keep a high level of code quality.

So why make the switch after all the praise?

I love programming. I especially love learning and playing with new languages. I love to dive in and explore and while sometimes this is beneficial (like stumbling across some new, weird framework written by some European), other times it is a serious distraction and causes a loss of focus. Feather became that distraction.

With Feather, ever time I went to fire off a blog post, I ended up working on the code base instead. For a while I was fine with it. Getting my hands dirty with Merb and contributing to an open source project was more important to me than developing an online presence. I like the people, I liked the code, and I was having fun.

But then a few weeks ago I was reading a post by Jay Fields and he was talking about being your startup. In the article Jay talks how important it is to develop your brand. He outlines a few points that he feels are important to accomplishing this. First on his list is having a blog.

At first I said to myself, “Sweet! I have a blog already! 37 Signals, are you ready to hire me?”, but then he made the obvious but overlooked point that you need to write and write often. This was where I was failing. I was using my blog to write code, not articles. Writing code isn’t bad, actually one of my favorite things to do in life really, but writing code when you aren’t suppose to be writing code is bad.

First you need to add a remote branch to your repository that points to the original repo you forked from.

git remote add --track master mleung git://github.com/mleung/feather.git

You will want to replace ‘master’ with the branch you want to track in the remote repo. In most cases this will be master, although you could replace it with edge or any other branch. You should also replace ‘mleung’ is what you the remote will be called.

To verify the remote repository was added run

git remote

You should see the new remote repo, in this case named ‘mleung’, along with any other remote repositories you may have previously added.

Now we can fetch all the changes from mleung’s code base.

git fetch mleung

This will create a new remote branch called ‘mleung/master’. Now we are ready to merge the code from the remote repository.

git merge mleung/master

That’s it. Remember, this process isn’t limited only to the original repository. Feel free to add remote branches for other user’s forks or even from repositories outside Github.

Hoptoad is a Rails plugin that acts a replacement for the exception_notifier from the wonderfully talented Thoughbot team (of Paperclip and Shoulda fame). Hoptoad is a hosted solution that keeps track of duplicate bugs, frequency, and resolution. This is a huge improvement over the current exception_notifier process which simply fires off an e-mail (or a few hundred while out of town one weekend). Instead, Hoptoad will differentiate new exceptions from reoccurring exceptions and allow you to track everything in a much more manageable way.

After attending a talk by Chuck Phillips on agile planning and estimation, I started looking for a more agile project management tool. I ended up going with Pivotal Tracker. Pivotal Tracker is a story-based project planning tool that gives teams a fluid platform for agile development. One of the coolest things about Pivotal Tracker is that it is velocity based meaning each story is assigned a point rating and then progress is estimate by historical velocity.

Have you recently added a new service or process to your workflow? Let me know in a comment.

Turns out there is a bug is merb-cache (part of merb-more) that is causing the problem. It’s patched in wycatz’ branch so pull down his latest version and get blogging!

Special thanks to my buddy Peter to figuring this one out.