I took this approach one step further and added some nice colored icons that represent the status of the current git repo:

This sort of functionality already exists of course, but I took this opportunity to learn more about shell scripting. So don’t use my solution as the best one out there, it is merely the result of me reading up on shell scripting. First make a copy of your ~/.bash_profile file. Paste ( and/or append ) the contents of this gist ( as seen below ) in your ~/.bash_profile ( create one if it doesn’t exist yet ), and open a new Terminal window to see the changes. Make sure you are in a folder that contains a git repo to actually see the branch name and colorful checkmark / x icons.

Homebrew makes it very easy to install the correct compiler, needed to build Ruby 1.9.2 and previous versions. Parallel to Xcode 4.3 and it's toolchain.
Check if you already have homebrew installed:
$ which brew


If all you get is an empty line or an error, you probably need to install Homebrew:
$ /usr/bin/ruby -e "$(/usr/bin/curl -fksSL https://raw.github.com/mxcl/homebrew/master/Library/Contributions/install_homebrew.rb)"


It doesn't hurt to check your ( existing ) Homebrew installation for possible problems.
The output from the doctor command is very helpful. Read it and act to it.
$ brew doctor


Make absolutely sure you have the latest and greatest:
$ brew update


Optional: handy tab-completion ( DO READ THE ADDITIONAL INFO after installing )
$ brew install bash-completion


Install RVM if you haven't done so already:
# RVM with latest, stable Ruby
$ curl -L get.rvm.io | bash -s stable --ruby


Using 1.9.3 as the default version of Ruby on your system:
# use 1.9.3 as the default Ruby version
$ rvm use 1.9.3-p194 --default


Handy alias pointing to 1.9.3:
# make an alias so you can use: `rvm use 1.9.3` ( without the patch number )
$ rvm alias create 1.9.3 1.9.3-p194


This is where the best part of the post is,
to install Ruby prior to 1.9.3 in combination with XCode 4.3+ on Lion.

Unlock extra formulas for homebrew:
$ brew tap homebrew/dupes


Installs gcc without messing with your Xcode compiler tools:
$ brew install apple-gcc42


Installs the older Ruby 1.8.7 patch 357: closest supported version.
$ rvm install 1.8.7


Installs Ruby 1.9.2-p320
$ rvm install 1.9.2-p320


Now run:
$ rvm list

If all goes well you should see something like:
ruby-1.8.7-p358 [ i686 ]
ruby-1.9.2-p320 [ x86_64 ]
=* ruby-1.9.3-p194 [ x86_64 ]


That's it. Depending on your machine, this should take approx. 30 minutes to complete every step in this post.

I wanted to load a little data into the test app so I decided to create a Sinatra app. An iOS app without some server interaction? Boring!

UPDATE: since I wrote this post, I found a better way. Read more here.

Using RVM I installed Ruby 1.9.2 and whipped up some classy code. Looking good, ok, let’s add some gems to the Gemfile, right, save, bundle install, crash – boom -bang …
Damn. Installing the gems failed.

Failed to build gem native extension.
Could not create Makefile due to some reason,
probably lack of necessary libraries and/or headers.

After a lot of trial & error, looking up possible problems, I ended up here.
The post states that in order to install Ruby 1.9.2 on OS X Lion 10.7.3 you should specify the compiler, overriding the default setting in the process.
rvm install ruby-1.9.2-p290 --with-gcc=clang

And indeed, this solved my problem.

bundle install started working as expected.

I want to give you a headstart with these books, all available on Safari ( O’Reilly’s online reading platform ).

• Away3D is a 3D engine for Flash
• AS3Mod is a cross-engine modifier library for Flash
• This explorer shows you how the output of your settings will look like and some might benefit from taking a look at the source code.

Away3D AS3Mod Explorer

1. One of the first things I did was “add new code snippet”. I immediately wanted to created my very own snippets folder, but couldn’t see anything that allowed me to do so. Too bad. I just created a new “code snippet” in stead. A popup jumps up, I saved my new fantastic piece of code and … nothing. The snippet wasn’t added anywhere. Opened all the default snippets folders, but my snippet was gone! Let’s try again. This time it worked. A “custom” folder was created in the code snippet panel and my code snippet was added as expected. Mmmm … not a great start, but maybe it’s just me.
2. So I clicked on the code snippet and it was inserted nicely in the actions layer of my project, but saw that something was wrong with my script. So I tried to right click the snippet and looked for an “edit code snippet” entry in the context menu. But I didn’t find such entry. I did, however, find a “edit code snippets xml”.
   
   
3. Ok, let’s do the “edit code snippets xml” action. Flash opens a new tab containing the xml for the code snippets. But half of my code is colored green. Correct XML color syntaxing anyone? Reason: I’ve used a single quote in the description of one of my snippets.
   
   
   Because the <description> tag doesn’t get CDATA from the code snippet panel, you are left out with just letters and some plain characters if you don’t want to ruin the color syntax in the XML editor tab. Booh!
   It’s also not possible to fix this by manually adding CDATA to the description tag. The code stays green.
4. A fourth very annoying thing is the lack of intendations in the “create new code snippet” popup. If you press tab on your keyboard, it just jumps around to the next UI component in the popup. That’s very nice, but not when you’re trying to right some reusable code!
   

This is where I end my rant on the new – not ready for production workflows – code snippets feature in Flash CS5. It is an great idea to help your fellow collegues with your collection of code snippets, but it lacks a lot of real usage testing in my very humble opinion. The Code Snippet feature in Flash CS5 is not worthy of your bucks until further development has been done.

Http://www.webdevotion.be/lab/screator/

Apparantly, Postgresql can’t convert a text column into a boolean, even when passing default values. The standard way of writing the migration didn’t work:
change_column :projects, :status, :boolean, :default=>true

So I started googling and ended up with:
execute "alter table projects ALTER COLUMN status TYPE boolean USING CASE status WHEN '1' THEN true ELSE false END;"

While this line did work, there is a problem with it; it’s not agnostic. It’s written specifically for Postgresql, but may cause issues in other environments.

So, the final, agnostic solution ( thanks Simon! ) uses plain Rails code. Basically we add a temporary column of type :boolean to store the information. After looping over all Project records we have converted all text values to the type boolean. Because we have all information stored in the temporary column, it’s ok to remove the existing status column and rename convert_status to status.

class ConvertStatusToBoolean < ActiveRecord::Migration
  def self.up
    add_column :projects, :convert_status, :boolean, :default => true
    
    # look up the schema's to be able to re-inspect the Project model
    # http://apidock.com/rails/ActiveRecord/Base/reset_column_information/class
    Project.reset_column_information
    
    # loop over the collection
    Project.all.each do |p|
        p.convert_status = p.status == '1'
        p.save
    end
    
    # remove the older status column
    remove_column :projects, :status
    # rename the convert_status to status column
    rename_column :projects,:convert_status,:status
  end

  def self.down
    change_column :projects, :status, :text
  end
end

Handy cheat sheet:
http://dizzy.co.uk/ruby_on_rails/cheatsheets/rails-migrations#execute

Useful info on Apidock:
http://apidock.com/rails/ActiveRecord/Base/reset_column_information/class

I’m not saying Ubuntu is not suited for more complicated user profiles. ( With the recent foul play of Apple, I would be more than inclined to use Ubuntu on my MBP on my day to day work. ) I’m just saying Ubuntu is easy enough to start with for most people to enter a virus free world. When they see the Firefox logo they now where “the internet” is.

But enough mumbo jumbo: how did the installation on the netbook go? Well, very smooth. No complaints. The installation was very straight forward. But… the graphics performance was bad. Problem: Intel’s GMA500 gfx card. Intel is not known as a member of the “we support Linux” pool. So, the community is basicly left on it’s own. Luckily for us there is a solution. Created by the community. It’s a work in progress, but for now it supports enough to work on your GMA powered netbook. The biggest thing missing is 3D support. But if you need that, why would you buy a netbook?

1. First check this wonderful Google Project
2. Then have a look at my xorg.conf file:

   Section "DRI"
   	Mode	0666
   EndSection
   
   Section "Extensions"
   	Option "Composite" "Enable"
   endSection
   
   Section "Device"
   	Identifier	"Configured Video Device"
   	Driver	"psb"
   	Option	"ShadowFB"	"True"
   EndSection

3. Check your grub file ( /etc/default/grub ) and DON’T forget to sudo run update-grub afterwards.

   # If you change this file, run 'update-grub' afterwards to update
   # /boot/grub/grub.cfg.
   
   GRUB_DEFAULT=0
   GRUB_HIDDEN_TIMEOUT=0
   GRUB_HIDDEN_TIMEOUT_QUIET=true
   GRUB_TIMEOUT=10
   GRUB_DISTRIBUTOR=`lsb_release -i -s 2> /dev/null || echo Debian`
   GRUB_CMDLINE_LINUX_DEFAULT="quiet acpi_osi=Linux acpi_backlight=vendor mem=2000mb splash"
   GRUB_CMDLINE_LINUX=""
   
   # Uncomment to disable graphical terminal (grub-pc only)
   #GRUB_TERMINAL=console
   
   # The resolution used on graphical terminal
   # note that you can use only modes which your graphic card supports via VBE
   # you can see them in real GRUB with the command `vbeinfo'
   #GRUB_GFXMODE=640x480
   
   # Uncomment if you don't want GRUB to pass "root=UUID=xxx" parameter to Linux
   #GRUB_DISABLE_LINUX_UUID=true
   
   # Uncomment to disable generation of recovery mode menu entries
   #GRUB_DISABLE_LINUX_RECOVERY="true"
   
   # Uncomment to get a beep at grub start
   #GRUB_INIT_TUNE="480 440 1"
   

4. Restart your system and your screen resolution should be ok. WIFI works as expected ( writing this blog post on the EEE PC now on WIFI ).

As I’m going to pass this system on shortly after this blog post I will not be able to check additional settings for you. All I can say is: I installed Ubuntu Lucid Netbook Edition from a USB I created on the existing Windows installation on this machine, before overwriting it with the Ubuntu installation. After that I saw that the screen reso wasn’t ok, so I installed the repo’s from the aforementioned Google Project page. After digging through some Googled pages I found the correct settings for my screen. It is my expectation that the people behind the Google Project will continue their good work and make a fully compatible driver available for us in the foreseeable future.

Apparat is a framework to optimize ABC, SWC and SWF files.

From the description of the Google Code page we get a hint of what Apparat actually does: “Apparat is a framework to optimize ABC, SWC and SWF files.”. In layman’s terms: “Apparat will speed up your Actionscript projects by optimizing certain method calls.”. The Actionscript Library for example provides some ( much ) faster ways of doing math and bitwise operations. Read on to understand more about how to get up and running with Apparat.

First, a confession. This post you are reading right now, is not the original one. After I thought I was done writing it, I sent it over to Joa and asked for a * small * review. A review turned into a complete rewrite as I missed an essential part of the project and made things more complicated for people who just want to integrate Apparat into their projects. So, credit where due: thank you Joa. Without his help this post wouldn’t be where it is now.

All in all, there are only three steps:

1. Download Apparat ( the compiled version )
2. Download Scala ( Apparat is ( being re-) written Scala )
3. Integrate Apparat with a project ( we’ll use Joa’s example )

Download Apparat

Let’s get started by downloading Apparat:

http://code.google.com/p/apparat/downloads/detail?name=apparat.zip

This file contains the compiled SWC among other compiled build tools ( Reducer, TDSI, Apparat, … ) that are part of Apparat.
You can extract this file wherever you want but it makes sense to put it somewhere close to your projects so you can go back to it easily in the future.

Download Scala

We should install Scala from http://www.scala-lang.org/downloads. Scala is a general purpose programming language and states that “Code sizes are typically reduced by a factor of two to three when compared to an equivalent Java application”. Sounds interesting to me, but even more to Joa I guess, as he is using Scala to get Apparat to the next level.

The version you need is RC3 at the time of writing this post. Windows users can download this zip, while the Mac / Unix archive would be this tgz archive.

You can put the extracted folder anywhere, you don’t need to install anything. It does makes sense to put it somewhere next to your working directories. We’ll need the Scala directory in a moment, but first we’ll set up our ( example ) project.

Integrate Apparat with a project

For brevity’s sake and to make this howto as easy to follow along as possible, we’ll download Joa’s most recent example:

get it from http://blog.joa-ebert.com/2010/05/26/new-apparat-example/.

After you extracted the source folder of the example project, you should put it in a nice place. Somewhere next to your other Flash projects maybe. Fire up FDT* and click File > New > New Flash Project from the application menu ( top of the window ) and point to the example folder as your working directory.

* We’re using FDT because it allready comes with an ANT builder plugin. It should be possible to use Flash Builder but you’re on your own ( for now – maybe I’ll write something about that in a next post if there is interest ).

If you’re not familiar with ANT you will learn something sweet from this post. ANT is a build tool, that can automate the sometimes tedious process of building a project. Imagine you want to compile an FLA from within FDT or Flash Builder, ANT can do that for you. It can even upload a compiled file to your staging server if you want. Do read more about ANT to learn all about it; its definitely worth it.

In the introduction of Joa’s blog post we read:

A complete example is also available. Just change the paths in the build.properties file and compile everything using Ant.

So let’s do that. Open up the example project in your FDT workspace and edit the build.properties file. You can find all files related to ANT tasks in the “build” directory. You’ll find three empty variables there: FLEX_HOME, SCALA_HOME and APPARAT_HOME. Add the correct values for your machine. Windows users should use \ ( backslashes ) in stead of the forward slashes I’m using on my OS X machine. Update: everyone ( Mac & Win ) should use forward slashes.

Save your changes in the build.properties file.

Next we’ll link our project to the compiled Apparat.swc file. The SWC is included in the Apparat download we extracted earlier. In the readme file of the example project we learn that we need to pass an APPARAT_FLEX path to our Flash project in FDT. Let’s do that by rightclicking on our example project and choosing New > Linked Libraries. Click the “add” button and create a new path “APPARAT_FLEX”. The path should point to your Apparat folder containing the Apparat.swc file.

Now we should be able to compile the project using the ANT task. Rightclick on the build/build.xml file. Choose Run as > Ant Build from the context menu. You should see ANT triggering several tasks in the Console view, with success off course. The result is a compiled version of both the example-apparat.swf and example-as3.swf files in your “bin” folder.

You can see a clear difference in the framerate of the two swf files. On my recent MBP I get 28fps on the AS3 example and a whopping 40fps on the Apparat version. And don’t forget: that’s 80.000 particles flying by in front of you!

If you see some errors in the Console, try to figure out what’s wrong, chances are one of the variables we edited before in the build.properties file are wrong.

“What’s the difference between me and you?”

What makes the Apparat version so fast? Inline methods sir. If you check out the two class files in the “src” folder of the example project in your FDT workspace, you can see they look almost the same. The only difference is that the the Apparat class file is using FastMath. As you can see from this screenshot ( click for full size ):

FastMath is a class provided by the Apparat SWC ( which we linked before ) and speeds up some mathematical functions. These function are recognized and “inlined”, making them so much faster. Inlining means that the compiler will put the actual method in the executed code, in stead of calling it from memory. So in stead of storing “Fastmath.rsqrt” it will put the actual code of the FastMath.rsqrt method where ever it is needed when compiling.

E.g.: in stead of using:

// AS3 version
dt = 1.0 / Math.sqrt(dx * dx + dy * dy)

the Apparat class uses:
// the faster Apparat way, will get inlined by the compiler
dt = FastMath.rsqrt(dx * dx + dy * dy)

Don’t forget to initialize FastMath when using it in your own projects!

Conclusion

Getting Apparat into your project should be very easy once you have gone through these ( actually simple ) steps. The learning curve was a bit steep for me, but I’m so glad I finally pulled it off with the help of Mr. Ebert. Please comment, ask, criticize in the comments below and tell us about your progress with this howto.

If there is interest in articles like these, I’ll try to post more Apparat related articles in the nearby future.