chatr - Tastes Great, Doesn't Leave a Bruise!

Craftsman Swap - Day 3

2010-01-21T00:00:00-08:00

Craftsman Swap - Day 3

21 Jan 2010

After a great first couple of days I was really starting to catch a groove at the Obtiva office. I switched off to a new project and a new pair. The task today was to take a look at integrating Paypal’s express checkout for an online retail application in Rails. My pair for the day was Nate Jackson. Nate started out at Obtiva as an apprentice and became a full time employee after his journey was complete. It’s great to see people who have enjoyed the experience and had such a positive view of their mentorship that they want to continue to work with their mentors. It is also great to see the mentors accept everyone as peers. It might not seem like a big deal, but it makes a huge difference!

One of the things I love about pairing is sitting down for the first time with somebody and observing how they work. Ruby on Rails gives us a common ground for ideas and communication, but the way we communicate is always different. It’s fun to see different tools and environments, and to see in what ways people invest in their tools to help make them a more productive programmer. We started out our pairing by firing up the one true editor and taking a tour of the application to get an idea of what we needed to do to accomplish our goals. The morning seemed to fly by. As soon as we were done walking through the application and reading the PayPal documentation it was time to head to lunch.

After lunch it was time to get cracking on our task. We were off to a great start with PayPal using the active_merchant gem. Unfortunately that didn’t last long. We had setup the basic developer sandbox accounts and got a simple example working, but found out that the proper solution was the “Pro” remix. This involved much more setup with the developer sandbox. The nice part is that since the Obtiva gang works together in the same room, we were able to get the pre-baked account credentials from others that knew where we needed to be. After fighting for a good part of the day around settings in the developer sandbox, and things not working properly, we arrived at what seemed to be a reasonable solution. The bad part was that PayPal just wasn’t cooperating with us. We were able to confirm that things were working with a live account, just not a developer sandbox account.

Nate and I were pleasantly suprised at the end of the day by a phone call from Carson Conant over at Medifly. Palm had contacted him to produce an app for the Pre and Pixi that would be able to stream the “Hope for Haiti” concert and allow for people to donate directly from their phones. The only downside was that we would only have one day to complete it. Obtiva had already built the original Pre app for Mediafly so we at least had a base app to work from. I was excited and nervous all at once for what day four would bring.

Craftsman Swap - Day 2

2010-01-20T00:00:00-08:00

Craftsman Swap - Day 2

20 Jan 2010

Context and situational awareness are a very important key to working well with others, and day two of my journey solidified that idea. The day started off similar to my first day. I was excited to be working with Ryan on Mad Mimi for another day. Having a better understanding of how Ryan works and how things at Obtiva flow during the day helped me concentrate more on getting real work done. After the morning standup Ryan and I were off to new feature development. We started off by writing tests for importing csv lists of recipients or “audience members” for email campaigns. Ryan had already done a lot of the leg work and refactoring, so it was all about tying a bow around things. Ryan and I were into a good pairing groove and the morning seemed to fly by.

Lunch time brought about a tradition that every company should adopt. Obtiva calls this tradition Geekfest, which is a “lunch and learn” style gathering where someone gives a talk on something new. I know that a lot of people do this informally, but Obtiva has made this a regular event and I applaud the effort that can go into to herding a bunch of developers in a room during the work week. I was also surprised to see the actual number of employees that Obtiva has. If it weren’t for Geekfest I probably wouldn’t have been able to meet all the other developers that work on site during the week. I gave a presentation on compojure, a Sinatra like web framework for clojure.

The day rounded out with Ryan and I making some UI enhancements and finishing up the feature from the morning. Since the deployment strategy involves multiple servers, keeping the csv lists around for batch processing presented an interesting deployment issue. We wrote out a few ideas for deployment solutions and then got in contact with Rails Machine to get their opinion so we could start prepping for the feature release.

Since this was my last day pairing with Ryan we ended the day with a retrospective to talk about how our pairing experience went. A few others sat in to offer outside perspectives and chat about initial impressions and what could be better throughout the rest of the week. This was a great way to end the day and pump me up for what’s to come!

Craftsman Swap - Day 1

2010-01-19T00:00:00-08:00

Craftsman Swap - Day 1

19 Jan 2010

I started my journey Sunday night to Chicago where I was welcomed into the Obtiva corporate apartment by Ethan Gunderson. I was impressed by the layout and attention to design. It is a swank little downtown pad that anyone would be happy to have. Ethan is one of Obtiva’s apprentices, working for 6 months at Obtiva where he contributes to the team and sharpens his skills as a developer. I really like this idea and would love to see more companies making good developers great.

Monday reminded me of the things I love about being in a big city. The walk to the office in the wonderful Chicago weather made for a nice wake up. The morning started out much like other good agile shops i’ve been to with a morning stand up meeting. A lot of the Obtiva are on site at clients so the meeting was very short. I noticed that after the company wide standup other teams of people had more focused standups which is a great idea. It let’s the group get to work while details are discussed among the relevant parties.

For my first day I got to pair with Ryan Briones. I know Ryan from my days in Columbus OH, so it was nice to start my week of with a familiar face. We started with a brief discussion about the project, Mad Mimi, which is an email marketing system. Ryan had plans laid out for the days work, but as things go we were able to defeat those plans and work on an entirely different set of problems.

We spent the first part of our day working on a little bit of housekeeping and deployment related issues. It was nice to see the server setup and hosting company choices. I also got to see the response and support from Rails Machine first hand as one of our deployments failed. Within a couple of minutes they were on top of everything before anyone could even ask. This kind of reaction to a problem combined with great support staff which they seem to have make for a great hosting experience.

After a great burger at lunch Ryan and I jumped back into things and were greeted with an interesting problem. A company was trying to start an email campaign with a record breaking amount of emails. Things weren’t going quite to plan and it was up to Ryan and I (with some help from Dave Hoover) to track down and fix the problem. After some debugging and extra logging (it’s like cowbell, you always need more), we were able to make some modifications and get the process moving. There are still a few performance tweaks to make, but history was made.

Once things were moving along nicely we moved into some other bug fixes to round out the day. We got to sit down this time and write some tests that proved our bug (and a couple other bugs that weren’t immediately apparent) and get to work. We were able to end the day by getting the bug squashed and watching the record breaking email campaign start it’s mailings.

The evening brought good food and good company as I went to dinner with Ryan, Ethan, and Ryan’s wife Stephanie. We ate some sushi down by Ryan’s house at a place called Ra. We ended the night with some coffee and good conversation that I hope will continue throughout the rest of the week.

Craftsman Swap

2010-01-17T00:00:00-08:00

Craftsman Swap

17 Jan 2010

I am headed to Chicago this week to work at Obtiva for the first of what I hope will be many craftsman swaps. I am really looking forward to hanging out with the crew up there and learning how other top notch development companies work. I will be keeping notes of my experiences while I am up there and will report a few times during the week.

Tell RCov to Cov Off

2009-10-16T00:00:00-07:00

Tell RCov to Cov Off

16 Oct 2009

Chad and I just released a new version of RCov! This version offers the ability to tell RCov to ignore certain pieces of code. This is useful if you are experiencing an bug in RCov, or if you just plain old disagree with RCov about the coverage status of a particular chunk of code. Here’s an example of implementation:

class Foo
  def bar
    1 + 2
  end

  #:nocov:
  def quux
    "s" + "hit"
  end
  #:nocov:

end

require 'test/unit'
class FooTest < Test::Unit::TestCase
  def test_bar
    assert_equal 3, Foo.new.bar
  end
end

Running this will show you 100% code coverage. This feature (for now) is pretty simple and does not do anything more than tell RCov to count the code as covered. If this feature takes off, I will spend more time making it smarter about the code and reporting so that you can easily tell the difference.

RCov Examples

2009-08-14T00:00:00-07:00

RCov Examples

14 Aug 2009

Are you interested in trying out RCov? Are you unsure of how to start? Are you looking to improve the RCov-fu on your project? I have just the solution for you! I have started a project that demonstrates different RCov configurations for different testing frameworks. This project is meant to help spread RCov related knowledge. You can get the bits by running:

git clone git://github.com/abedra/rcov-examples.git

Feel free to browse around and submit any feedback. I am hoping to get as many samples in as possible, so feel free to contribute!

Getting Started With Compojure

2009-08-11T00:00:00-07:00

Getting Started With Compojure

11 Aug 2009

Compojure is a great web framework written in Clojure. Although in its infancy, the project has a lot of promise. The best part is that getting started is a snap!

The only thing you need to have installed on your system is Java and Ant. Everything else is taken care of for you by Compojure. To make things even simpler I have provided a script that does all of the project setup for you. Let’s take a look.

After you have downloaded the script, make sure to make it executable (chmod +x compojure) then run it:

compojure <project name>

where project name is whatever you want it to be. The script will download Compojure and it’s dependencies, do a fresh build, and copy all of the necessary dependencies for your project into the directory you specified. It will then leave you with a <project name>.clj file that has a “Hello World” example built right in. When the script finishes running just run:

./start

and go to http://localhost:8080. You are now up and running on Compojure! To get a better idea of how to work with Compojure you can check out this site.

Failing Your Builds with RCov

2009-08-07T00:00:00-07:00

Failing Your Builds with RCov

07 Aug 2009

I just pushed the latest bits to RCov, and they include a new option to help you fail in more spectacular ways! Have you ever made sure that your build fails if you don’t satisfy the coverage threshold? If you are in RSpec land, you have always been able to use the built in verify_rcov task, but if you are using any of the other testing frameworks, it has historically been quite a process.

To use the new hotness, simply add --failure-threshold xx where xx is the threshold you want your coverage to be at. If your coverage drops below that threshold you will get a message telling you, but more importantly your build will actually fail. RCov exits non-zero when you fail to satisfy the coverage threshold which means that your Continuous Integration builds will fail. Having this in your bag of tricks is an awesome thing, and can help ensure that everyone is checking in tests along with their code.

To get the latest RCov bits simply run:

gem install relevance-rcov -s http://gems.github.com

You should get at least version 0.8.5. The --failure-threshold option will be included in RCov 0.8.5 and higher. As always, if you have any issues please report them to the RCov issue tracker and I will be sure to take a look. If you are going to submit a bug, check out this tasty groove that will give you some tips on how to properly submit a bug report for RCov. Enjoy your Whales!

RubyNation Wrapup

2009-06-13T00:00:00-07:00

RubyNation Wrapup

13 Jun 2009

RubyNation has been a very nice conference. The quality of talks was excellent as well as the crowd. I got a chance to meet some very cool people and even talk a little bit of Clojure. The slides for my talk on RCov and Ruby 1.9 can be found here.

Getting RunCodeRun Build Status for a Specific Project

2009-06-05T00:00:00-07:00

Getting RunCodeRun Build Status for a Specific Project

05 Jun 2009

Are you on RunCodeRun yet? If not, stop, go to RunCodeRun and signup! Now that you have the awesome sauce I am going to show you how to display RunCodeRun build statuses for individual projects anywhere on your website. Earlier this year Glenn Vanderburg created the RunCodeRun Badge Project that allows you to display your RunCodeRun build statues on any website. This project is very cool and gives you a nice piece of art for your site. This project is a modification of Dr. Nic’s Github Badge Project. The focus of this article is to arm you with the codez to have individual control over where and how you display RunCodeRun build statuses. If you look at the landing page for my blog you will notice that I have this hotness already implemented.

It turns out that this is incredibly simple to do. Let’s take a look at the basics…

$.getScript("http://runcoderun.com/api/v1/json/abedra?callback=parseData")

This starts all the magic. It calls out to the RunCodeRun API and returns a JSON payload with all the open source build statuses linked to your account. Let’s break this down quick. $.getScript() is a jQuery function that allows you to do an Ajax request to a different domain than the one your site is currently running on. This is important because what you are essentially doing is a Cross Site Scripting exploit on your own site. This, however, is acceptable to do if you know what is actually happening. The reason a standard jQuery $.get won’t work is because it violates Javascript’s same origin policy constraints. The next thing to break down here is the ?callback=parseData at the end of the request. This is done so that the data we get back from the $.getScript() call can be processed. Under the covers the parseData() function looks like this:

function parseData(data) {
  var castronaut = data.user.projects[0];
  var rcov = data.user.projects[4];
  var safe_erb = data.user.projects[3];
  var mpi_ruby = data.user.projects[2];

  $("#rcov").append(generateStatus(rcov));
  $("#castronaut").append(generateStatus(castronaut));
  $("#mpi-ruby").append(generateStatus(mpi_ruby));
  $("#safe-erb").append(generateStatus(safe_erb));
}

A user object is returned to you that contains an array of projects. For now, the RunCodeRun API does not have a named key for each project, so you have to use the appropriate array index for each project in order to get the details. parseData() calls the generateStatus() function:

function generateStatus(project) {
  if(project.status == "success")
    return "<span class='success'> succeeded </span>" + "<span>" + prettyDate(project.ended_at) + " <a href=" + project.url + ">details</a></span>";
  else
    return "<span class='failure'> failed </span>" + "<span>" + prettyDate(project.ended_at) + " <a href=" + project.url + ">details</a></span>";
}

generateStatus() then calls prettyDate(). This is a small library written by John Resig. I had to make a small patch to get it to work with the dates that the RunCodeRun API sends back. The patched version can be found here.

Takeaway

The full implementation exists in the source of this site at http://github.com/abedra/abedra.github.com. If you just want something simple to display the status of all your builds use the RunCodeRun Badge Project. If you want a little more control over what is displayed this should get you off to a good start.

DC Developer Day Wrapup

2009-05-30T00:00:00-07:00

DC Developer Day Wrapup

30 May 2009

The DC Developer Day was a hit! There were a lot of good talks and interesting people. For a closer look at the events that transpired you can search twitter for #devdaydc. You can find my slides on RCov and the Ruby 1.9 changes here.

Castronaut Updates - Version 0.7.5 Released

2009-05-22T00:00:00-07:00

Castronaut Updates - Version 0.7.5 Released

22 May 2009

Castronaut has had some much needed lovin applied to it. I have updated the test suite to work with the latest version of RSpec. The test suite is happy again and all of the deprication warnings left over from the Sinatra upgrade are now resolved. The gemspec has been cleaned to remove some unnecessary files and an XSS bug has been fixed that was present on the login page. Just a reminder that issue tracking has been switched to Github Issues, so if you encounter any problems please post them over at http://github.com/relevance/castronaut/issues.

Central Ohio ISSA Information Security Summit Wrapup

2009-05-16T00:00:00-07:00

Central Ohio ISSA Information Security Summit Wrapup

16 May 2009

The Infosec Summit was a great success! There were a lot of great talks, and some pretty cool demonstrations. I am happy to have been a part of it and hope to head back next year. If you missed it, you should definitely put it on your conference watch list. Slides from my talk can be found HERE.

Laziness, I Haz It

2009-04-23T00:00:00-07:00

Laziness, I Haz It

23 Apr 2009

It’s been a while since I posted anything at all. I apologize for all the radio silence, and hope to do better from now on. In all of my laziness over the past few months I decided to move by blog over to github and take the github pages way out. This is going to make things much simpler to manage, since I use the incredibly awesome Github to manage my code I thought it would be best just to keep things all in the same place.

I do have some cool new stuff in the works, so stay tuned for updates on my recent talks, ideas, and open source projects.

Tarantula 0.0.8.1 Released

2009-01-16T00:00:00-08:00

Tarantula 0.0.8.1 Released

16 Jan 2009

Jason did some great work today testing and cleaning up some of the rough edges with Tarantula. A new version has been released and it’s now getting ever so close to hitting 1.0. You can install the gem from github via

% gem install relevance-tarantula --source http://gems.github.com

As always, if you find any bugs or are having trouble with Tarantula you can report them at the Tarantula Lighthouse.

Stop. Emacs Time.

2008-10-03T00:00:00-07:00

Stop. Emacs Time.

03 Oct 2008

I decided to get all crazy and re-theme my blog based on my favorite editor. It’s still got a couple of rough edges, but it’s coming along. Please comment with any strangeness.

Windy City Rails Wrap-up

2008-09-23T00:00:00-07:00

Windy City Rails Wrap-up

23 Sep 2008

Saturday’s Windy City Rails event was great! There were some great talks and good conversation. I would like to thank Chicago Ruby along with the conference organizers who made all of this possible.

You can download slides for my Sleight of Hand and my Security talk in the previous links.

Speaking at Windy City Rails

2008-08-20T00:00:00-07:00

Speaking at Windy City Rails

20 Aug 2008

I just got word from the tribal council at the Windy City Rails conference that they have accepted two of my proposals. I will be giving my RailsConf 2008 talk “De-Railing: Smashing the Rails Stack” along with “Sleight of Hand for the Ruby Man”. The De-Railing talk is a security focused session on steps to take to ensure that your Rails application has what it needs to keep data safe from attackers. The Sleight of Hand talk is a session on metaprogramming tips, tricks, and pitfalls with Ruby and with Rails. Stay tuned for more info closer to the September 20th conference date.

Adventures in Clojure - Getting Started

2008-08-17T00:00:00-07:00

Adventures in Clojure - Getting Started

17 Aug 2008

Following Stu’s series on Java.next I decided to take a look into Clojure. Clojure is a pretty cool language. It’s a lisp dialect that runs on the JVM. This is cool for a couple of reasons. First, you get all the advantages of the jvm. Second, you get the advantage of interoperability with java libraries. I am not necessarily excited about the later, nonetheless, there are still advantages to having solutions to problems that you can use to quickly get things done.

Although clojure has a nice website with decent language descriptions and a helpful Wiki, getting started can leave you a little wanting. Here’s a more precise setup guide to get you going. I am using Ubuntu Hardy on my desktop PC so this guide will be geared towards it.

First, let’s install the necessary base libraries

% sudo apt-get install sun-java6-jdk maven2 subversion

Now let’s check out the clojure source and build it

% svn co https://clojure.svn.sourceforge.net/svnroot/clojure/trunk clojure
% cd clojure
% mvn install

When maven gets finished you should end up with a shiny new jar in the target dir of the clojure source directory. Now to test this out let’s try the following from inside the clojure directory.

% java -cp target/clojure-lang-1.0-SNAPSHOT.jar clojure.lang.Repl
Clojure
user=>

Now that we have the basics setup we need to get our environment setup for real programming. The first enhancements come to the repl. It would be nice to have command completion, parenthesis matching (this is lisp after all), history across sessions, and last but not least editor hooks. The wiki suggests both JLine and rlwrap. JLine is ok, but lacking in the ability to do command completion and some other goodies. That being said let’s install rlwrap so that we get the most out of our environment.

% sudo apt-get install rlwrap

Next we need to create a script to make launching the clojure environment a little bit easier. Here’a a script from the wiki to help us out.

#!/bin/bash
BREAK_CHARS="(){}[],^%$#@\"\";:''|\\"
CLOJURE_DIR=/usr/local/clojure
CLOJURE_JAR=$CLOJURE_DIR/target/clojure-lang-1.0-SNAPSHOT.jar
if [ -z "$1" ]; then
    rlwrap --remember -c -b $BREAK_CHARS -f /home/abedra/.clj_completions \
        java -cp $CLOJURE_JAR clojure.lang.Repl
else
    java -cp $CLOJURE_JAR clojure.lang.Script $1
fi

As you can see I moved the clojure directory to /usr/local/. I also had to modify the path to the .clj_completions file. Just copy this script, save it, chmod +x, and move it to /usr/local/bin/. Make sure you properly modify the paths. Last, just go ahead a create an empty file at ~/.clj_completions. We will populate that next. Let’s use this script to populate the .clj_completions.

(defmacro with-out-file [pathname & body]
  `(with-open stream# (new java.io.FileWriter ~pathname)
     (binding [*out* stream#]
       ~@body)))

(def completions (keys (ns-publics (find-ns 'clojure))))
(with-out-file "clj-keys.txt" (doseq x completions (println x)))

Save this code as clj-completions.clj. Now simply run the script.

% clj clj-completions.clj

Once it’s finished you’ll end up with a file named clj-keys.txt. Simply move that file to ~/.clj_completions and you are ready to rock. Now let’s test out our new hotness. Once you invoke clj try typing in some code and pressing tab.

Now you are ready to get programming! Stay tuned for more adventures in clojure.

Upgrades

2008-07-09T00:00:00-07:00

Upgrades

09 Jul 2008

I just updated to the current git master version of mephisto. I was running an old version of 7.3, so if anything is out of whack please shoot me a message and let me know. I am working on getting the syntax highlighting working again, it was suffering from some boundary issues.

Sleight of Hand for the Ruby Man

2008-06-12T00:00:00-07:00

Sleight of Hand for the Ruby Man

12 Jun 2008

Chad Humphries pointed me to this tasty bit of code. Since Ruby so graciously lets you open up classes anywhere, it’s nice to know where the right place to debug is. Simply put the following code into a globally accessible place (I just used .irbrc) and you will have the method available.

module Kernel
  def where_is_this_defined(settings={}, &block)
    settings[:debug] ||= false
    settings[:educated_guess] ||= false

    events = []

    set_trace_func lambda { |event, file, line, id, binding, classname|
    events << { :event => event, :file => file, :line => line, :id => id, :binding => binding, :classname => classname }
 
    if settings[:debug]
      puts "event => #{event}"
      puts "file => #{file}"
      puts "line => #{line}"
      puts "id => #{id}"
      puts "binding => #{binding}"
      puts "classname => #{classname}"
      puts ''
    end 
    }
    yield
    set_trace_func(nil)

    events.each do |event|
      next unless event[:event] == 'call' or (event[:event] == 'return' and event[:classname].included_modules.include?(ActiveRecord::Associations))
      return "#{event[:classname]} received message '#{event[:id]}', Line \##{event[:line]} of #{event[:file]}"
    end

    if settings[:educated_guess] and events.size > 3
      event = events[-3]
      return "#{event[:classname]} received message '#{event[:id]}', Line \##{event[:line]} of #{event[:file]}"
    end

    return 'Unable to determine where method was defined.'
  end
end

Once you have that you can simply open up script/console or irb and give it a whirl

>> where_is_this_defined {Streamlined.ui_for(:foo)}
=> "Streamlined received message 'ui_for', Line #6 of /Users/abedra/src/someapp/trunk/lib/extensions/streamlined/streamlined.rb"

This tells us that the ui_for method of streamlined is actually being called from an extension that the project made rather than the streamlined plugin itself. This would save you lots of time trying to debug the wrong method! Give it a try.

Rails on OpenBSD 4.3

2008-06-07T00:00:00-07:00

Rails on OpenBSD 4.3

07 Jun 2008

OpenBSD released version 4.3 of their OS on May first. This version is especially exciting because there has been considerable effort made by the ports contributors to get the Rails stack up to date. I will run through a full stack setup from a scratch install. If you need information on installing OpenBSD there is a great tutorial that will get you started.

Post Install

You will be greeted with a vanilla install and not a whole lot to work with. Let’s get some basic sytem tools installed. We will use OpenBSD’s pkg_add tool to get us up and running quickly. The stack we will setup will include helpful system tools, Ruby / Rails, Thin, MySQL, and nginx.

Helpful System Tools

# build the locate database
% /usr/libexec/locate.updatedb
# setup pkg_add path for easy installs
% export PKG_PATH=ftp://ftp.openbsd.org/pub/OpenBSD/4.3/packages/i386/
# install bash (optional)
% pkg_add -i bash
% chsh -s bash
# install other needed system tools.  we use pkg_add with the -i
# flag for interactive mode.  this way it will ask you what version
# of a specific software you want if there is more than one available.
% pkg_add -i curl nano subversion git

Ruby / Rails / Thin

% pkg_add -i ruby ruby-iconv ruby-gems
% gem update --system
% gem install rails thin
# if you want to use thin with sockets you will need to install
# eventmachine from source to get version 0.11
% gem install eventmachine --source http://code.macournoyer.com

MySQL

% pkg_add -i mysql-server
% mysql_install_db
# start mysql
% mysqld_safe &

Create the file /etc/rc.conf.local and place the following into it.

mysql="YES"

Now open up /etc/rc.local and add the following script just before the ending echo ”.”.

if [ X"${mysql}" == X"YES" -a -x /usr/local/bin/mysqld_safe ]; then

    echo -n " mysqld"; /usr/local/bin/mysqld_safe --user=_mysql --log --open-files-limit=256 &

    for i in 1 2 3 4 5 6; do
        if [ -S /var/run/mysql/mysql.sock ]; then
            break
        else
            sleep 1
            echo -n "."
        fi
    done
fi

This will setup mysql to start when the system boots.

nginx

% pkg_add -i nginx

Then open up /etc/rc.local and add the following after your mysql entry.

if [ -x /usr/local/sbin/nginx ]; then
   echo -n ' nginx'; /usr/local/sbin/nginx
fi

This will setup nginx to start when the system boots. I like to modify my nginx configs a little bit taking a little away from the debian linux apache2 and nginx configs.

% cd /etc/nginx
% mv nginx.conf nginx.conf.orig
% mkdir sites-available
% mkdir sites-enabled
% curl -O http://aaronbedra.com/nginx.conf
% cd sites-available
% curl -O http://aaronbedra.com/vhost

This will enable nginx to automagically load any vhost file that is in sites-enabled. We use the sites-available and sites-enabled directories and symlink from one to another. I prefer this, you don’t have to do this if you already have an nginx setup you like to use. Assuming you followed the above steps the only thing you have left to do is edit the vhost file and add your sites information. I would suggest renaming the vhost file to something a little more descriptive as well. Don’t forget to create the directory in /var/log/nginx/yoursite or nginx will fail to start. Once you have completed that you just need to symlink your new config into sites-enabled and restart nginx.

% ln -s /etc/nginx/sites-available/vhost /etc/nginx/sites-enabled/vhost
# start nginx
% nginx
# restart nginx
% kill -HUP `cat /var/run/nginx.pid`

Post Setup Notes

There is one small problem with eventmachine on OpenBSD. When thin intializes eventmachine at application boot time, it tries to use the wrong eventmachine library. You need to use the pure ruby version. An easy way to fix this is to add the following to your .profile

export EVENTMACHINE_LIBRARY="pure_ruby"

This will make eventmachine happy and thin able to run. The reason for choosing thin and nginx over the other implementations is because thin can connect to to nginx via sockets, and nginx is just plain easier to configure, not to mention faster. Obviously you can season this setup to taste if you prefer other stack implementations.

Surviving the Reboot

It’s good practice to have a thin config file that can start your application. Let’s take a look at how to configure thin for your app.

% mkdir /etc/thin
% thin config -C /etc/thin/your_config.yml -s2 -S /tmp/thin.sock

This will create a thin config file that will spawn to instances of thin (s2) and use sockets instead of ports. You can change the number of instances it spawns or where you want to put your sockets if you so desire. Now that you have a running system, you just need to add one more thing to make sure that your app starts if your server reboots. To make this work you need to do two things.

% ln -s /usr/local/bin/thin /usr/bin/thin

This will make sure thin is available at boot time. Next you need to add a line in etc/rc.local to start your app. Add this after the other startup scripts you added previously.

echo -n " thin"; thin start -C /etc/thin/your_config.yml

Happy OpenBSD-ing!

Railsconf Talk

2008-06-01T00:00:00-07:00

Railsconf Talk

01 Jun 2008

Here are the slides from my railsconf talk. Thank you to all of you that showed up and gave support!

Shame on you O'Reilly!

2008-05-30T00:00:00-07:00

Shame on you O’Reilly!

30 May 2008

I have a bone to pick with the mensa that decided that it was important to give each Railsconf attendee a badge that had all of their personal information encoded on a barcode on the back side.

From the back of the badge:

“Your name, email address, job title, organization, address and phone number are encoded on this card. Please recycle this card at the end of the conference”

What? Is there a magical card swiper that gives you money and ever lasting happiness? Exactly why do I need to carry around a badge with all that information? Do they really expect me to give it back to someone else for “recycling” at the end of the conference? This is one badge that is going in the shredder!