The Irish Penguin Watching Open Source unfold across Ireland

21Jun/100

Deploying your Qt on Rails apps on the N900 (Maemo)

Qt on Rails is a framework to let you turn your Rails sites in desktop applications and harness the power of Ruby! It's not at production level yet but it's certainly possible to have a good play with it and a bit of a hack! If you're not familiar with Qt on Rails then a good place to start is this blog post covering the v0.1 release. Also, check out the github repo for more info on installing Qt on Rails on your desktop and building an application with it. Here we show you how to deploy Qt on Rails based apps on your N900. One of the goals of Qt on Rails is to provide an easy way for you to develop apps faster for Maemo and, down the road, hopefully MeeGo too!

Note: This blog post may help you figure out how to install any QtRuby application on the N900, not just Qt on Rails apps. Also, this QtRuby Maemo wiki article was particularly useful when I was stumbling along this path!

One thing you will need to install as part of this guide is Easy Debian. Easy Debian greatly expands what you can do with your Maemo device. It basically sticks a full-featured version of Debian on your device. This means 2 things - firstly, for the uber-geeks out there it let's you fire up a Linux desktop on the N900; though it's important to note that your normal Maemo desktop isn't affected by Easy Debian. Secondly, having a full-on Debian available let's you run Linux apps such as Open Office! Sweet! And what rocks is that you can even run these programs without invoking the Easy Debian Linux desktop - in a seamless manner. It's important to note that the user interface to these Easy Debian-based apps behave a differently to a typical native Maemo program; rather they work like a traditional desktop application with a mouse pointer on screen.

Install Steps
  • Note: For simplicity, this guide assumes you are installing an application which stores data using sqlite3. Also, the steps here have been tested against the N900 firmware update PR1.2. If you are using an older version of the firmware you may want to consider updating it.
  • Firstly, install Easy Debian with the N900's Application Manager
  • Install the Easy Debian image via the new Deb Img Install application added to your list of applications
    • Note: This is a 1 gig download, but comes with cool stuff like OpenOffice and intergrates pretty seamlessly with your desktop
    • Takes an hour or so to download and then extract itself
  • Open the Debian Chroot terminal (not the usual N900 terminal), which should now be in your list of applications
  • Install rubygems, qtruby and  sqlite3 with ruby bindings
    • sudo apt-get install rubygems
    • sudo apt-get install libqt4-ruby
    • apt-get install libsqlite3-ruby
  • Install the bits we need need from Rails (without installing documentation)
    • sudo gem install activerecord activesupport activeresource --no-ri --no-rdoc
  • Zip up your Qt on Rails application and copy to any directory on to the N900. Note that the Qt on Rails application consists of the entire Rails directory directory including the vendor/plugins/qtonrails directory intact and  a sqlite3 database already created under the db directory).

    If you don't have your own Qt on Rails application to hand then you can create the RadRadio app discussed in the "Make it so, Jim!" section of the v0.1 release blog post

    In the Qt on Rails v0.1 release there is a bug that accidentally introduces a dependency on a korundum library, which is not needed in this case. An issue is logged against this in the Qt on Rails Issue Tracker As a workaround, find and remove any occurrences of require 'korundum4' in files under the vendor/plugins/qtonrails/app directory

  • Once transferred, simply unzip it on your device. Note: If you saved the zip to the Documents folder on your N900, this can be found under /home/user/MyDocs/.documents when poking around the filesystem
  • Finally, via the good ol' Debian Chroot terminal, change directory to the vendor/plugins/qtonrails directory of your app and execute

    ruby1.8 run

  • Boom! You should see your Qt on Rails app in all it's glory!

Note there is a bug where you cannot input into a text field when running a Qt on Rails app on the N900 using above technique (seamless mode). As a workaround, open the Qt on Rails app inside of the Debian LXDE desktop (rather than in seamless mode). You can find Debian LXDE in the list of applications on your device. Inside Debian LXDE, open a terminal and run the application as above. Just a quick heads up, sticky keys don't work like you'd expect - you have to hold down the Shift and Fn keys to use them.

21Jun/102

Qt on Rails v0.1 released. But is this Ruby-based Qt and KDE app framework doomed?

Ruby has changed the way developers build web applications. Since popularised by the Rails framework, programmers no longer stumble around in the dark with disparate web forms; instead they are able to clearly focus on the business problem and expose a well-modelled domain in a easily testable manner. Traditionally used in data-heavy domains, today's web apps now encroach on the desktop's home turf of rich highly-functional applications - something years ago thought impossible. And most surprisingly, through clever use of patterns and conventions, they've arguably become the easier of the two to develop. Given this, could desktop developers learn from the web app approach? This is in part the motivation behind Qt on Rails - let's use a conventions-based approach to building desktop applications and clients. Let's harness the ease and expressiveness of the Ruby language. And let's have a clean consistent domain model underneath the hood with a comprehensive suite of tests to boot. A grand idea; but right now, it's on course to fail...

List Page - Disc Jockeys
Main window listing Disc Jockeys (Click the image to enlarge)

What does this framework look like?

What exactly is Qt on Rails? Well first, let me just make a little guilty admission. This blog post is aimed at Ruby and/or Rails developers first, then Qt/KDE developers second. This is not because I believe one camp is more important than the other. It is because I really want to bring Rails dev's on-board to KDE/Qt development and I see a real need to give them a first-class Qt toolkit; to make desktop apps as brilliant and easy to develop as their web apps. If Qt on Rails ever makes it as far as being a fully mature framework, I hope that a Rails developer using it for the first time will find it a very familiar experience. The directory structure, naming conventions and overall architecture (to date at least) has been chosen to that end. But I hope to do an article focusing on approaching this project from a more on Qt/KDE perspective. So what is Qt on Rails like?

Well, imagine you just wrote a Rails web application. You're finished! Let's say we just built a web app for a fictional company called RAD Radio. We have two important things in our system - Disc Jockeys and Sponsors for the radio shows. So we have this really neat model layer sitting on top of our database. It handles all our business logic and things like validation of data being persisted - let's reuse this... verbatim! This is our line in the sand! Design choice 1, Qt on Rails literally reuses everything from the model layer down already within your Rails application!

If we start from the front end of our proposed desktop application and work backwards, we want to have a Qt app which consists of various different screens, some of which may be on display simultaneously. This is a little different from the web, because generally on the web you can think of having just one screen and this gets blown away on each request (unless AJAX intervenes, but still you get the overall point). In a Qt desktop app we decide that our initial screen is to be "main" window for our application. Clicking on a button may cause a new window to launch, for example, a window to edit a Disc Jockey's details, but the main window will remain there, though not active. This kind of difference (from your typical web behaviour) presents a difficult challenge and will have a big effect on our architecture - for example, it has influenced my decision not to try and reuse anything for the controller layer generated in a Rail's application. Back to our story, we have a main window as our starting point, which you can see in the figure above. It should have a way to navigate between different parts of your application. This is achieved through the big command link buttons at the top of the window. In our web application, each part of our application is based on different sources of data (called resources when using RESTful terminology). Clicking on the Sponsors command link button would cause the grid to be refreshed with Sponsor records. And then clicking on the Disc Jockeys command link button refresh the grid with Disc Jockey records again.

This sounds reasonable so far. A quick side note though; bear in mind this is still early days for Qt on Rails. I'm sure many good KDE and Qt developers may be horrified at the UI decisions made above. I'm not a Qt or KDE expert! This is another big challenge. One of the most crucial areas we need help on is getting feedback on how Qt and KDE applications are generally laid out, what widgets are used for what purposes and so on. In essence, what are the best practice guidelines for HCI in something like KDE and how can we incorporate that into the apps we generate with Qt on Rails.

So we are looking at the main window, which contains a list of Disc Jockeys in a grid - how does this work? Well, first we need to be aware of what Qt is giving us here. The list widget that you see on screen is a Qt widget called a QTableView. It is concerned only with displaying stuff. Underneath that we have a QAbstractTableModel. This holds the data, it does not care about how it will be displayed. Now it's worth pointing out that a QAbstractTableModel has nothing to do with the Rails concept of a model. You see, the Qt folks have used the MVC pattern in their architecture for donkey's years now; and it's important to not confuse the two different uses of the same paradigm. The QAbstractTableModel is an object into which you stick a whole bunch of Rails records that you wish to display in a list. We then plug the QAbstractTableModel into the QTableView widget and that's how your records are displayed. It's also important to note that when a QTableView sees the QAbstractTableModel being plugged in - it has absolutely no idea that the underlying records are of type Disc Jockey. Think of the QAbstractTableModel as an adapter between a collection of records for a particular Rails model and the QTableView widget which will ultimately display them. Probably most importantly, remember that

  • A Rail's model equates to one record
  • Whereas a QAbstractTableModel equates to many records - it sits on top of a collection of Rails records and allows them to be displayed in something like a QTableView widget

Note: The Qt API is incredible. It's extremely comprehensive! Check it out at http://doc.qt.nokia.com You will want to look at the "C++ Application Development Framework" for the version of Qt you develop with. One big tip to note down - where it talks about something like a QAbstractTableModel in the docs - is that it is referring to the C++ world. In your head just translate this to Qt::TableModel - now you can happily use all the documentation available. Also, in your code you will always write Qt::AbstractTableModel, never QAbstractTableModel.

So to edit a Disc Jockey record I select a row and click the Edit button near the bottom of the page. Hey presto! An Edit Disc Jockey form appears...

Edit Form - Disc Jockey
Form for editing a Disc Jockey (Click the image to enlarge)

Nice! And when you click the Save button the Edit form is dismissed and the list of Disc Jockeys in our main window is refreshed!

Cool! But where do all the Qt on Rails files live?

Qt on Rails is installed in the vendor/plugins directory of your Rails app (see http://github.com/theirishpenguin/qtonrails for more details on how to install). Under vendor/plugins/qtonrails/ there is a directory called app/ which holds the Qt application code. In turn, app/ is divided into the following subdirectories, with each subdirectory corresponding to a layer in our application framework, listed here from the highest level (like the HTML stuff in a web app) to the lowest (the controller in our case - as everything from models down to db is handled by vanilla Rails).

  • qdesigns - XML markup files (with a .ui extension) which are to Qt screens/widgets what HTML files are to HTML pages/elements
  • ui_proxies - A Ruby representation of .ui files - this is an autogenerated layer which you don't really need to worry about
  • qpresenters - Where the presentation logic (not any business logic) for your screen lives
  • qhelpers - A place to put logic that you wish to reuse across presenters
  • qviews - Where we decide exactly how we will build a view for a particular controller action (I accept that the name 'qviews' is confusing; maybe 'qviewbuilders' would be better or something totally different)
  • qcontrollers - Where we decide what data to retrieve for a particular action (though we don't specify 'where to go' in the way a Rails controller has render() or redirect_to() methods)
I am the Executioner...

So here's the flow of control. In our vender/plugins/qtonrails directory we have a 'run' command. When we execute it, the command...

  1. Fires up a Qt Application instance
  2. Asks our (very primitive) Router to take a given route and instantiate a QView and a QController for it
  3. The QController then fetches data depending on the action (just like a Rails action) and hands off the data to the QView
  4. The QView then decides what screen should be build (or whether to stay on the same screen and perhaps just refresh a list of records)
  5. Once the appropriate Screen has been built, it is displayed with a show() call. At this point the user will then see something happen on-screen.

The above is pretty much like one request-response cycle in a Rails app. Now we play the waiting game

  • The user does some*thing like selects a row and clicks the Edit button.
  • At this point the QPresenter comes into play. The QPresenter is your window widget. I didn't tell you then, but the QPresenter was created earlier in the QView layer when we call the constructor for the window (ie. the QPresenter) we wish to display). I'm open to the fact that QPresenter may not be the best name for this! Anyway, the QPresenter contains your presentation logic. In our case, the user has just clicked the Edit button on the MainWindow presenter. This causes the handler for the event - the edit_clicked() slot to be called. Qt uses a brilliant concept called Signals and Slots to handle events in your application - a Signal is something that acts as a trigger (such as a button being clicked) for a Slot; a Slot simply being a function dedicated to doing something useful in response. A Signal is connected to a Slot using the aptly named connect() method. I neglected to bore you with this little detail earlier but this connection was carried out in the constructor for the QPresenter (ie. the window) which took place a few steps back when we mentioned that the QView layer decided what screen is to be built
  • Finally, the slot - edit_clicked() in this case - asks the Router to take a given route and now we're back to step 2 of the flow of control outlined earlier
  • Phew!
Make it so, Jim!

So that's it in a nutshell. How do get all this lovely goodness to turn your Rails web app into a skeleton desktop app? One command, mon ami! From the root of your Rails directory simply run

./script/generate qtify DiscJockey Sponsor

So let's say you have no Rails app right now. Here's how to get to a basic web app and skeleton desktop app in double quick time!

Firstly, just a quick note on OS dependencies. Qt on Rails has mainly been developed to date on Kubuntu 9.04, Kubuntu 9.10 and the Ubuntu Netbook Remix 9.10. For these platforms, you can install the packages mentioned in the "First Install the appropriate packages" section of this Developing Qt4 Applications using Qt Designer and Ruby on Kubuntu article, which also goes into more detail on QtRuby development if you're interested. This QtRuby bindings article on the KDE TechBase also gives some useful info on the Ruby bindings for Qt.

We've not tried this yet on Windows or Mac, but here's a Windows QtRuby install guide by Richard Dale and a QtRuby on the Mac install guide. So you are welcome to try and install QtRuby on one of these platforms but we can't promise anything!

Perhaps, most excitingly of all, Qt on Rails apps can be deployed to Maemo devices such as the N900! Check out this related Deploying your Qt on Rails apps on the N900 (Maemo) article! And MeeGo support will surely follow sometime soon!

Ok, let's cook you up a Rails app. Here we are using Rails 2.3.5 and assuming you are at a Linux command line (see Qt on Rails github project page for more installation details of Qt on Rails itself)...

rails RadRadio

cd RadRadio

./script/plugin install git://github.com/theirishpenguin/qtonrails.git

./script/generate scaffold DiscJockey name:string date_of_birth:date programme_name:string radio_slot:time max_num_guests:integer next_review:datetime

./script/generate scaffold Sponsors name:string signed_up_on:date

rake db:migrate

./script/generate qtify DiscJockey Sponsor

cd vendor/plugins/qtonrails

./run

# If ./run on it's own gave you a permission error you can try 'ruby ./run' instead

# Oh yeah!

What you see is almost what you get...

Hopefully, you should be seeing an app around about now. Once you have some rows in the list of disc jockeys or sponsors (create some DJs using the New button), you will see that cells of the grid that make up the list can be edited in-place. This is quite a powerful feature to have out of the box. An Edit button is provided also, though if you plan to provide an Edit button to your users, which launches a form, then you should probably make the grid read-only so as not to confuse them by having two ways of editing. As this is a early prototype of the framework, I've left the Edit button and in-place editing enabled, trusting you to tailor them to your app's needs. If you were going to switch off in-place editing you would probably also want to select the entire row (rather than just one cell) when you click on a cell in the row.

Let's have a quick look at form validation in action. Add a validates_presence_of validator to your DiscJockey and Sponsor Rails models so that they look as follows

#In app/models/disc_jockey.rb (under root of your Rails app, not under qtonrails)

class DiscJockey < ActiveRecord::Base
validates_presence_of :name
end

#In app/models/sponsor.rb (under root of your Rails app, not under qtonrails)

class Sponsor < ActiveRecord::Base
validates_presence_of :name
end

Close and restart the RadRadio Qt on Rails app. When you hit the New button and try to create a new Disc Jockey or Sponsor without a name, you will see that the Rails model validation kicks in and you get a validation message telling you that you need to correct the name field. Validation also works if you are editing a record after clicking the Edit button, however validation messages don't appear if you edit a record in-place in the grid (just because I haven't had a chance to implement that yet).

Again, as it's an early release there's also plenty of bugs in there, such as a crash if you try to click the Edit Button without selecting any row. A list of issues is maintained on the Qt on Rails Issue Tracker. Check it out to see what limitations currently exist and add to it if you spot a new problem!

Doomed?

As you can see, this codebase is being opened up quite early. A good start has been made - but the project is still very embryonic! Surely, it's a bit early to be talking about doom? Well, unless hacking on Qt on Rails appeals to some developers out there it will certainly die a merry death on the great scrapheap of nice projects that just didn't make it. Why? Not only is there a lot of work that needs to be done, but also it's more fun to work on the project when there's a crew involved, which also brings new ideas and energy to any project. Otherwise it's unsustainable.

But life wouldn't be fun it it wasn't a bit of a challenge, right? If you think this kind of project might interest you, if you're not put off by insurmountable odds, then know that your framework needs you! Drop me an email at declan [[AT]] weuseopensource [[DOT]] com or twit a quick tweet on twitter (theirishpenguin). The Qt on Rails Issue Tracker is a good place to start looking for things to hack on. Or you're welcome to fork the project on github and develop that killer feature you'd like to see in there!

Qt on Rails, because it's at an early stage, is an easy place to start - there isn't lots of code to weigh you down. The framework itself is very thin, the majority of the code is actually Rails style generators to take the Rails model layer and build the Qt/KDE app on top of it. Because of the tiny codebase, it shouldn't be hard too get your head around what's going on.

Qt on Rails doesn't want to be doomed!

It's intended that Qt on Rails gives Ruby and/or Rails developers an outlet to develop first class desktop apps using the best available framework. Rails developers often ask, "If I want to build a great cross-platform desktop app, what GUI toolkit would I use?" The answer varies and there no one headline GUI toolkit/desktop framework that currently has mindshare amongst Ruby developers and works great across multiple platforms. Given that Qt is currently so strong across the Linux desktop, commercial Windows applications, Macs and mobile platforms such as Maemo and MeeGo why shouldn't it be the first thing you reach for when building a Ruby desktop app? Come help us build Qt on Rails! Stave off the doom!

26Apr/100

Using git-svn to connect to a Subversion repository via Git

This is quite simple on Linux.
* Install git, subversion and git-svn (the latter can be installed on Ubuntu using apt-get install git-svn)
* Acquire a copy of an svn repository (or just a folder in the repository) as a git repository locally
git svn clone https://yoursubversionserver.com/svn/trunk/some_folder
Note: You can choose to omit the 'some_folder' at the end if you want everything in trunk. You can also be more specific about which folder you are interested in, eg.
git svn clone https://yoursubversionserver.com/svn/trunk/some_folder/wow/really/specifc/folder
* Change a file in the repository and use the usual git commit command to commit it to a local repository
* To push changes from your local git repo to the subversion repo use
git svn dcommit
* To pull down changes from the remote subversion repo to your local git repository use
git svn rebase
* Just to be clear - in the last step we really did use the git svn rebase command to get changes from the remote repository.

** Do not use git pull. See the REBASE VS. PULL/MERGE section of the git-svn docs as to why **

** Do not use git clone/pull/merge/push on your local git repositories that are derived from a subversion repo. See the CAVEATS section of the git-svn docs as to why **

More basic examples are also list at the bottom of the git-svn docs.

And if you want to see a great example of a git-svn workflow in action, check out Jérémie Laval's blog post - Working on Mono with git-svn.

16Apr/103

Lifting the lid on Open Jam!

Open Jam was an Ubuntu Ireland lead event which invited all members of the Open Source community to come along to Enterprise Ireland's Dublin offices at East Point on Saturday the 27th of March. And come along they surely did, really show-casing the diversity of groups we have here in Ireland - users of Open Source software, developers, admins and advocates. The timing of Open Jam was to coincide with the Ubuntu Global Jam - where contributors to the Ubuntu project focus on finding, prioritising and fixing bugs as well improving documentation, artwork and more. The 'Open' in Open Jam was to extend that spirit to everyone - independent of their area of interest or skillset - to collaborate, learn and share.

After the welcoming talk, some people got straight down to business, while others took the opportunity to chat and chill out after a long week in the office. Thanks to the excellent organisation skills of David Scanlon of Enterprise Ireland the event sported two rooms - one main room where you could hack away on your favourite project and another where liked-minded folks could get together and have a BoF (Birds of a Feather) session on a particular topic. Tunes were supplied on the day, pumped through Enterprise Ireland's very tasty A/V system, by Luis Bethencourt - a.k.a. Mr Ubuntu Studio (Ubuntu Studio is a multimedia creation flavor of Ubuntu that Luis leads). In the BoF room, Rory Geoghegan kicked off the proceedings with a great intro to the Python programming language. This was followed by Anton Krasovsky's highly popular talk on Erlang - which seemed to bring out the inner developer in everyone and sparked a meandering conversation through programming patterns, architectures and practices of every language under the sun. Anton is the author of pavo.me - a mulitmedia Twitter client for any pretty much phone that supports Java.

After lunch, which was graciously sponsored by Microsoft, Rory McCann gave an insightful talk on Ubuntu's software collaboration platform Launchpad. Laurent Coudeur, part of the GNOME translations team, presented on the theme of language translations. He's always on the lookout for people to help the translation effort (especially regarding Irish, so if you're interested in helping then get in touch with him on his blog, where he also has a write up and more photos of Open Jam). Away from the lightning talks, Halo Labs, which is a community of Independent IT Service Providers, were beavering away on Linux appliances like Workgroup Server & Asterisk Box, with Patrick O'Conner and Russell Davies probably scooping the award for being the most industrious attendees!

Musical talent was a feature of the day, with Harry van Haaren from U.L. providing a really interesting lightning talk on his work in progress Luppp project; hooking up a MIDI board to his C++ based looping software - allowing part of a live instrument performance to be looped on the fly and built upon. UCD Open Source Labs (represented by Alexander Ufimtsev, Keith Byrne, Aidan Church, David Murphy and Chris Duffin) announced their availability to to support Open Source projects by providing facilities for collaboration, project planning and development of Open Source software - check out the UCD Open Source Labs website for more details.

The Open Source .NET scene had a strong presence with Dublin Alt.NET members on hand ready to participate in any conversation that had the mere mention of the words 'design pattern' :-) Andrea Magnorsky introduced the audience to the .NET DLR (Dynamic Language Runtime), which allows languages such as Ruby and Python to run on the .NET framework. Mono hackers Alan McGovern and Jérémie Laval gave a lively talk on the Mono framework, MonoDevelop and related platforms such as MonoTouch for the iPhone and Moonlight. Qamir Hussain presented on A.I. & Distributed Agent frameworks which he's working with at the moment. Later in the day, Rory McCann was back on stage, this time with Larry O'Neill, to give a talk on the facinating Open Street Map project, which marches on mapping the world in an Open fashion. Oh and some random walked in off the street and chipped in with a quick talk on Qt, Ruby and Rails :-)

To round off, again a big thanks to David at E.I. for being such a kind host and also Laura Czajkowski, Jeffrey Roe and Qamir Hussain for helping to organise the event. It's great to have Enterprise Ireland helping events like this, which really boost the Open Source community and innovation in Ireland. Most of all, thanks to you, everyone who showed up and helped make it a great event. The good news is that there's no let up in the flood of Open Source events in the next few weeks with OSSBarcamp on April 17th and Open Spaces Coding Day on April 24th. Phew! Being Open was never so easy!

P.S. While the resulting projects need not be Open Source, a superb sounding collaborative event that's on the horizon and worth a mention is Dublin's Startup Weekend! This could be your chance to get that great idea in the back of your head implemented over a weekend sprint with other techies involved. It's being organised by Sean Murphy and if interested visit the Startup Weekend website for more details.

P.S. If you want to leave your twitter handle to get in contact with others at the event then feel free to edit it into the Open Jam Attendees List

And finally, some more Open Jam pics...

4Feb/101

A Ruby Module that mixes in Class Methods (static) and Instance Methods

Ho ho, this one can catch you out more than once so it's high time to write a blog post to cover this off. Turns out it's a commonly used pattern to the rescue. Thanks to eoin on #ruby.ie for pointing to the solution on RailsTips.org.

Here's quite a tasty diagram too for easy reference.

module Swingable

    def self.included(base)
        base.extend(ClassMethods)
    end

    def instance_swing
        puts 'Did an instance swing!'
    end

    module ClassMethods
        def static_swing
            puts 'Did a static swing!'
        end
    end
end

class BaseballBat
   include Swingable
end

BaseballBat.static_swing
BaseballBat.new.instance_swing
18Dec/090

Santa’s got Gems baby! Ruby Ireland Christmas Meetup 2009

Ho ho ho! The month's Ruby Ireland meetup sprag right out of the traps with early adopters showing up at 6pm in the lobby area of the Trinity Capital Hotel, Wed Dec 16th. Easing into the evening with a 4 euro pint and talk of Android phones - seemingly the top item of everyone's Christmas shopping list - the latest crop of gems in the Ruby world was in hot debate, gemcutter in particular.

A couple of folks had been playing around with RubyGame for visualising data as it changes on the fly - showing that this framework is for more than just gaming. The XML/HTML parser Nokogiri was also mentioned a few of times in passing, with the particularly eye-catching quote "XML is like violence - if it doesn’t solve your problems, you are not using enough of it" adorning the home page of its website. And the cracking little tool tig was also brought up, which has a dinky little ncurses interface into git repositories. Pretty cool; not least because it makes it easier for newbies to avoid being bitten when they start git'tin.

The downstairs lobby in the hotel worked out great for people to meet up and relax, with most people turning up at the scheduled 7 o clock for kick off. From there we took over the, what has to be said, pretty classy meeting room complete with old style couches and some Joan Miró paintings. Just in tune with the creative buzz we had going on. There wasn't too much talk of Ruby for a while as most people were in stunned admiration of the room. Then the food platter arrived. Impressively, this is when everyone showed off their good manners by looking shyly at the platter for a few minutes, with that kind of "You first, sir" glint in their eye, before taking the plunge and sinking into the pakoras and wedges! Pretty much undoing any good work in the gym from earlier in the day!

One of the funnier moments of the night was when someone went to check the tweets against the (now settled upon) #rubyireland hashtag. Only to find lost rubyists tweeting from the hotel lobby as to where the meetup was on. After a quick runaround the lobby to herd anyone wielding a Macbook into the meeting room, the evening was back on track. We split up into a few smaller groups, with the main walk-through being on the qtonrails - a Rails plugin to simply developing applications on Linux and other platforms using Nokia's Qt framework atop Rails.

To finish off we had a bit of improv comedy from everyone at different closing stages of the evening; in particular Paul O'Malley with his faithful rendition of an emotion beekeeper. And yes now we're straying off topic so it's probably time to go. We'll leave you with Paul's write up of last night's shenanigans :-)

Thanks to everyone who showed. Have a great Christmas and catch ye all in Jan 2010 - surely destined to be the decade of Ruby domination!

Ciao,
Dec

26Nov/091

Generate Rails Migrations from your PostgreSQL or MySQL database

1) Create a new empty Rails project called schemer

2) In your config/database.yml file, point at the database you wish to dump to a migrations file

3) Run the command 'rake db:schema:dump'. This should create a db/schema.rb file. Amazingly this effectively is your migrations file!

4) To tidy up create a file called file db/migrate/20091125205635_create_initial_schema.rb

5) Then copy the create_table statements from the schema.rb file into the new file 20091125205635_create_initial_schema.rb. Here's a template

class CreateInitialSchema < ActiveRecord::Migration

  def self.up
    # Put all create_table statements from schema.rb file here
    # Note: You don't need the 'ActiveRecord::Schema.define(:version' line or it's enclosing end statement
    # ...
    # ...
  end

  def self.down
    # Don't really need this
  end

end

6) Once you've all this done you can just run 'rake db:migrate' and you should have a new sqlite db up and running under db/development.sqlite3

Thanks to Justin Ball on this Nobody Listens Anyway blog at Dump an Existing Database Schema Into a Ruby On Rails Migration Ready Format for the basis of this tip. Sometimes somebody does...

22Nov/094

Eight Useful Git Tips

1.) You can check your config using git config -l

This tells you useful stuff like what remotes you're bound to, etc.

2.) You can blow away files listed as 'untracked' (by the git status command) using git clean (be careful!).

If you want any of those files to hang around (but not appear as 'untracked') then add them to your  .gitignore file.

3.) You can re-enter your last commit using git commit --amend. What's nifty is that you can change files (including file addition and deletion) as well as the commit message. Just do what you gotta do before running the amend command and enter a replacement commit message.

4.) You can undo your last commit completely using git reset HEAD^

*** This does not change the working tree ***

Alternatively, you can use the syntax git HEAD~2 instead of HEAD^^

5.) Refer to the last revision by HEAD, the second last as HEAD^ and the third last as HEAD^^

You can keep adding carrets forever! Sounding like Bugs Bunny there.

6.) You can see the contents of a file from a revision using

git show rev:path/to/file

7.) Ok, this tip is a bit of a longer one. Use git merge some_other_branch to merge another branch into the one your working on. (Note that git pull does a merge automatically once it's completely - if you don't want this to happen use git fetch). There are a few possible outcomes:

Fast-forward merge : If changes were made on only one of the branches since the last merge, then the merge should happen without any need for you to act.

Three-way merge: If changes were made on both branches, then git decides how to merge them using an algorithm. If any changes conflict with each other, then git will report them and let you resolve them. Once you've fixed any conflicts then you can git commit.

If there were no conflicts, git automatically does a commit with a stock commit message for the log. If you don't like the idea of this then use git merge --no-commit some_other_branch. You need to manually do the commit afterwards. (This is more like how other distributed VCS's such as Mercurial. This was the tip that inspired this post!).

8.) Finally, one tip that I found useful is that you don't want to git pull into your repository when you have uncommitted changes. Especially if you treasure your sanity. Instead git stash is your friend. Here's how
* Do a git stash to move your uncommitted changes to a magical place in git temporarily
* Then do the git pull
* And finally do a git stash apply. This will now move your 'stashed' uncommitted changes back (from the magical temporary place in git) into your working directory and take into account what what was changed by the pull. Neat!

Tips 1-7 taken from:  Git - SVN Crash Course. Tip 8 taken from painful experience :-)

If you want a good detailed yet friendly reference on a particular command search the Git User Manual on kernel.org.

19Nov/090

Installing Ruby On Rails on Windows

There were a couple of great outcomes from the first Free Ruby Lesson we ran in the Havana cafe on Dublin's Georges St last Monday. The first was getting everyone hacking with the Rails stack and some practical examples in double quick time. The second was managing to get RoR installed on Windows Vista as the lesson rumbled on in the background. Here's how Rails conquered Vista.

We kicked off by following the instructions on the rails wiki. It's worth paying careful attention to anywhere it says to set path variables. If you find that you are getting an error that says that you version of rubygems is not recent enough (and you can't get rubygems to update itself) then you can install an older version of Rails, for example 2.3.2, using the following from the command line

gem install rails -v 2.3.2

(Note: If you have already installed a different version of rails you can uninstall it first by using 'gem uninstall rails').

The big problem however was getting sqlite3 working. We did install the SQLite Command Line Tool and the SQLite DLL as the Rails Wiki instructions said to - but to no avail. We kept getting a popup error saying that the sqlite3 dll was not found and to please consider reinstalling. Fortunately, we gambled on one of the answers on StackOverflow. Basically the solution was from the command line

gem uninstall sqlite3-ruby
gem install sqlite3-ruby --source http://gems.rubyinstaller.org

And with that we were away! The final thing was to get a nice friendly IDE installed so we plumbed for Netbeans on the Netbeans Downloads page. Look out for the special Ruby version of Netbeans which is one of the links around the middle of the downloads page.

Anyway, that's all for this week. Happy Ruby hacking!

29Oct/090

Understanding how Ruby stores objects in memory – the Ruby Heap

Ruby has it's own heap management which actually consists of several 'Ruby Heaps' to manage objects created during the execution of a Ruby program; this is separate from the System Heap for your Operating System. Each individual Ruby Heap contains Slots, with each Slot able to one reference one object.

The entire space that an object takes up in memory ***is not stored inside the Slot***. Rather each Slot is a small fixed size space which can be thought of as the Ruby interpreter's handle a location in memory. This location exists outside of the Ruby Heap itself and contains the real 'meat' of the object. To be clear, if you have a 50MB string - the 50MB of data is stored outside of Ruby's Heap. If you really want to know the story of the 50MB, the space for it is actually allocated by something like the malloc command in C (as good ol' Ruby is written in C) and then stored on the System Heap. The Slot in the Ruby Heap simply contains a reference to that memory location on the System Heap which contains the 50MB of data.

Here's an example. Let's say that a Ruby program creates a single string of 50MB
* A single free Slot in a Ruby Heap becomes filled
* Memory to store the 50MB of data that makes up the string itself is allocated in memory and put on the System Heap (outside the Ruby Heap!) and a reference to this location is stored in the Filled Slot on the Ruby Heap
* There comes a point in time when this string is no longer needed. This slot is garbage collected on the next GC iteration
* The Filled Slot is turned into a free slot. The 50MB of data in memory referred to by the slot is also freed and returned to the Operating System

Ruby starts of with a minimal set of Ruby Heaps. These are managed by by a Ruby Heap list. Ruby creates Ruby Heaps when needed and frees Ruby Heaps back to the OS when no longer needed (the latter is done in a sub-optimal manner - more on this later). Each Ruby Heap created will be 1.8 times the size of the previous heap. In other words, it will contain 1.8 times the number of slots of the previous heap. Ruby's Garbage Collector, periodically iterates through the Ruby Heaps and frees up any Slots as appropriate (and also the memory that an object really occupies which is referenced by the Slot - ie. the 50MB data of the String) back to the system. Once a GC iteration is complete, some of the Slots that were filled will now be empty - known as Free Slots. Remember that we said that Ruby's Heap management actually consists of many Ruby Heaps. Well if one of these Ruby Heaps consists of only Free Slots then the Ruby Heap itself will be freed back to the Operating System.

There is a problem with this last statement however - if a Ruby Heap contains mostly Free Slots and one Filled Slot then it will not be freed. You could have many Ruby Heaps in this state. As long as a Ruby Heap contains even one Filled Slot it will not be returned to the Operating System. It just takes one bad apple to spoil everything! What would be nice is if some sort of Heap Compaction (kind of like disk fragmentation) took place where all Filled slots were pushed together into completed filled Ruby Heaps. This would leave you with completely filled Ruby Heaps, one semi-filled Ruby Heap and then a bunch of completely empty Ruby Heaps. The completely empty Ruby Heaps could then be freed, releasing precious memory back to the Operating System. Alas the current mainstream Ruby interpreter does not do this.

References
* How the Ruby Heap is Implemented Phusion Passenger's Hong Lai gives a great explanation of the Ruby Heap - the banner may not be quite suitable for work. Fortunately, there's a censor button :-)

* Fine tuning your garbage collector Chris Heald explains some of the settings around garbage collection

* Ruby's Garbage Collections effect on Ruby on Rails Pluron Inc's blog discusses so of the knock-on effects of Ruby GC on Rails and importantly mentions the 8 MB memory allocation tigger for the garbage collector