juretta.com (Rails)

Goodby Rails, kind of...

Fri May 29 00:00:00 +1200 2009

I started this blog about 4 years ago using a custom built Rails blog engine using Rails 0.8. A Rails based blog engine wasn't around at that time and I wanted to build something using Rails anyway.

Having a custom built blog engine has its benefits and drawbacks. The rather basic administration interface was a pain to use and running two Mongrel instances for a nearly static website seemed to be a misallocation of ressources anyway.

So instead of switching to another full-fledged blog engine I switched to using Jekyll to generate a static version of the blog.

Tom Preston-Werner (the author of Jekyll) writes about his reasons for creating Jekkyl in Blogging Like a Hacker:

I was tired of complicated blogging engines like WordPress and Mephisto. I wanted to write great posts, not style a zillion template pages, moderate comments all day long, and constantly lag behind the latest software release.

Jekyll is the complete opposite, it is a:

[...] simple, blog aware, static site generator. It takes a template directory (representing the raw form of a website), runs it through Textile and Liquid converters, and spits out a complete, static website suitable for serving with Apache or your favorite web server.

With a basic directory layout and a simple configuration file creating a static copy of your blog is merely running the jekyll command in your blog directory.

stefan@macbook-pro:~/webs/juretta.com/www [0]$ tree -L 3
.
|-- Rakefile
|-- _config.yml
|-- site
|   |-- 404.html
|   |-- _includes
|   |   `-- entry-preview.html
|   |-- _layouts
|   |   |-- default.html
|   |   `-- post.html
|   |-- _posts
|   |   |-- 2008-11-11-hudson_test_drive_part_1_rails.html
|   |   |-- 2008-12-11-hudson_at_wellrailed.html
|   |   |-- 2009-05-24-activemq-jms-stomp.markdown
|   |   `-- 2009-05-29-goodbye-rails.markdown
|   |-- about
|   |   `-- index.html
|   |-- atom.xml
|   |-- favicon.ico
|   |-- images
|   |-- index.html
|   |-- javascripts
|   |-- rss.xml
|   |-- sitemap.xml
|   `-- stylesheets
`-- web

Running jekyll in this directory will copy all the static files in site into web and will transform your content files (HTML, Markdown, Textile based) applying a series of layout files (Read more on the Jekyll Wiki pages: Usage).

Running jekyll --server will transform the files and fires up a local WEBrick server running on http://localhost:4000.

stefan@macbook-pro:~/webs/juretta.com/www (master) [0]$ rake server
(in /Users/stefan/webs/juretta.com/www)
juicer merge --force --document-root ./site ./site/stylesheets/base.css
Produced site/stylesheets/base.min.css from
  site/stylesheets/undohtml.css
  site/stylesheets/screen.css
  site/stylesheets/form.css
  site/stylesheets/syntax.css
  site/stylesheets/base.css
jekyll --server
Configuration from ./_config.yml
Building site: site -> web
Successfully generated site: site -> web
[2009-05-29 17:49:10] INFO  WEBrick 1.3.1
[2009-05-29 17:49:10] INFO  ruby 1.8.6 (2007-09-24) [i686-darwin9.2.0]
[2009-05-29 17:49:10] INFO  WEBrick::HTTPServer#start: pid=7634 port=4000

Throw in some juicer goodness and a nice git post-receive hook and your new post is only a single git push away.

Hudson at WellRailed

Thu Dec 11 00:00:00 +1300 2008

I had the pleasure to give a short talk about Rails and Hudson at the Wellington Rails user group (WellRailed) meeting in November.

I talked about using the Hudson extensible continuous integration engine to run Rails tests and using git as the SCM.

The slides are availble here: static.coravy.com/talks/WellRailed-Hudson-Rails-git.pdf

Hudson test drive (Part 1: Rails)

Tue Nov 11 00:00:00 +1300 2008

I've been using Hudson for quite a while now and it just rocks. Gone are my CruiseControl (for Java) and CruiseControl.rb (for Ruby/Rails) installations. I even ditched most of my crontab entries and have Hudson manage the repeated tasks for me.

Hudson monitors executions of repeated jobs, such as building a software project or jobs run by cron. Among those things, current Hudson focuses on the following two jobs: 1. Building/testing software projects continuously just like CruiseControl or DamageControl. [...] 2. Monitoring executions of externally-run jobs, such as cron jobs and procmail jobs, even those that are run on a remote machine. [...]

Compared to CruiseControl configuring jobs in Hudson is really easy and straightforward. You can easily give it a test ride without even installing Hudson in your servlet container. Let's give it a try.

stefan@macbook-pro:~$ curl -OL http://hudson.gotdns.com/latest/hudson.war
  % Total    % Received % Xferd  
                                 
100 19.6M  100 19.6M    0     0  
stefan@macbook-pro:~$ java -jar hudson.war 
[Winstone 2008/10/24 12:33:11] - Beginning extraction from war file
hudson home directory: /Users/stefan/.hudson

This will start an instance of hudson which can be accessed at http://localhost:8080/

Hudson will create a .hudson directory to store files and plugins. You can set the HUDSON_HOME environment variable to use a different directory.

stefan@macbook-pro:~$ export HUDSON_HOME=/opt/local/hudson

Due to its multi purpose nature it's easy to build different kinds of software projects. If you can build your projetc using a command line script you might be able to build it using Hudson.

Hudson and Rails

Part one covers building Rails projects with Hudson. I'm going to cover building Java and Flex projects in future posts.

In order to build a Rails project as soon as a commit is made, we need a couple of plugins. I'm going to use the git scm for this example. The default Hudson installation supports SVN out of the box but with the Hudson git plugin using git as your scm is a piece of cake.

Go to Manage hudson (http://localhost:8080/manage) > Manage plugins (http://localhost:8080/pluginManager/) and click on the Available tab. Choose the following plugins:

Git Plugin
Rake plugin
Ruby Plugin

and click on the Install Button (bottom right).

Hudson needs to be restarted in order to activate the plugins. Stop the currently running proceess (CTRL+C) and start Hudson again (java -jar hudson.war).

Configure your Rails project

Hudson can run your Rake test task without further modification. But it would be really nice if the test output could be aggregated in Hudson.

We are going to install the ci_reporter gem (http://caldersphere.rubyforge.org/ci_reporter/) for this purpose.

CI::Reporter is an add-on to Test::Unit and RSpec that allows you to generate XML reports of your test and/or spec runs. The resulting files can be read by a continuous integration system that understands Ant‘s JUnit report XML format, thus allowing your CI system to track test/spec successes and failures.

stefan@macbook-pro:~$ sudo gem install ci_reporter
Successfully installed ci_reporter-1.5.1
1 gem installed
Installing ri documentation for ci_reporter-1.5.1...
Installing RDoc documentation for ci_reporter-1.5.1...

Add the following lines to your Rakefile (or clone the git repository with the Rails example project:
git clone http://github.com/juretta/hudson-example-project-rails/tree/master).

require 'rubygems'
# Install:
#  sudo gem install ci_reporter
gem 'ci_reporter'
require 'ci/reporter/rake/test_unit' # use this if you're using Test::Unit

Run the following rake task and the ci_reporter gem should create xml files in the results directory of your project.

stefan@macbook-pro:~$ rake ci:setup:testunit test:units \
    test:functionals CI_REPORTS=results

Add a new job

Go to the Hudson homepage and click on New Job (http://localhost:8080/newJob). Choose a Job name, select Build a free-style software project and click Ok.

You should now be on the project configuration page (http://localhost:8080/job/YOUR-PROJECT-NAME/configure).

In the Source Code Management section choose git. Enter your git repository URL (The Rails project used in this example is hosted on github: git://github.com/juretta/hudson-example-project-rails.git).
Click on Add build step and choose Invoke Rake. Enter db:migrate ci:setup:testunit test:units test:functionals CI_REPORTS=results RAILS_ENV=test into the Task field.
In the Post-build Actions section check Publish JUnit test result report and enter results/*.xml into the Test report XMLs field.
Click Save

You can click Build now to trigger the first build of your Rails project.

We now want to trigger the build automatically as soon as we do a git commit. Edit the post-receive file in the hooks directory of your git repository (.git/hooks/post-receive).

URL='http://localhost:8080/job/YOUR-PROJECT-NAME/build'

echo "Run Hudson build"
wget $URL > /dev/null 2>&1

Make the file executable.

stefan@macbook-pro:~$ chmod u+x .git/hooks/post-receive

The hook script now triggers a build after each commit.

The Ruby metrics plugin

The Ruby metrics plugin shows Rails stats (LOC) and rcov reports. Unfortunately with the latest version of Hudson the Ruby metrics Plugin does not seem to work.
You should definitely give it a try. If you get an HTTP 500 Status code with a org.apache.commons.jelly.JellyTagException
remove (or disable) the following line:
<st:include it="${it.owner}" page="sidepanel.jelly" /> in
$HUDSON_HOME/plugins/rubyMetrics/WEB-INF/classes/hudson/plugins/rubyMetrics/railsStats/RailsStatsBuildAction/index.jelly

Have fun with Hudson!

Rails WDDX request

Wed Jan 31 00:00:00 +1300 2007

I was asked how to get the body of an HTTP post request in a Rails controller. And with the most things in Rails it's very easy.

In Java you can get a ServletInputStream from the request object, in PHP there is $_SERVER['RAW_POST_DATA']. Pretty much like in the java version you can just ask the request object for raw_post in Rails.

# Simple RPC-Controller in Rails
# request.raw_post contains the raw post data, that is
# the data that is submitted in the http request body.

# The example uses the wddx gem to return a simple WDDX response message

# To us WDDX install the wddx gem ([sudo] gem install wddx)
# and a require "wddx" statement
# to config/environment.rb

# app/controllers/rpc_controller.rb

# This controller has a simple method "endpoint"
# The body of the post request is written to the 
# log/#{environment}.log file.
# The controller returns a WDDX XML message.
class RpcController < ApplicationController             
  # simple endpoint
  def endpoint
     if request.post?
       logger.debug request.raw_post
       render :xml => "Ok".to_wddx
     else
       render :xml => "Failure".to_wddx
     end
  end
end

To test this controller you can use the following code to issue a post request.

#!/usr/bin/env ruby
#
#  Created by Stefan Saasen on 2007-01-31.
#  Copyright (c) 2007. All rights reserved.
require 'rubygems'  
require 'net/http'
require 'wddx'

begin      
  http = Net::HTTP.new("localhost", 3000) 
  response = http.post('/Rpc/endpoint', {"key" => "value", 
      "another" => "value"}.to_wddx) 
  puts response.body
rescue SocketError => e
  # error handling
end

Rspec Rails plugin

Thu Dec 07 00:00:00 +1300 2006

I'm using the very convenient RSpec Rails plugin. From the Rspec website:

RSpec is a framework for practicing Behaviour Driven Development (BDD) in Ruby.

You can find more information on Rspec and the rails plugin on http://rspec.rubyforge.org/.

After using rspec for a little while the following error suddenly occured:

vendor/plugins/rspec/lib/spec/rails/version.rb:14: (RuntimeError) 
############################################################################
Your RSpec on Rails plugin is incompatible with your installed RSpec.

RSpec          : 0.7.4 (r1201)
RSpec on Rails : r35

Make sure your RSpec on Rails plugin is compatible with your RSpec gem.
See http://rspec.rubyforge.org/documentation/rails/install.html for details.
############################################################################

The RSpec plugin uses the Subversion Revision information to determine the current version number. If you are using Subversion to manage your rails project this error will occur if you have svn:keywords enabled and svn:keywords contains "Revision". The version information then reflects the revision of your repository.

You either have to disable svn:keywords for vendor/plugins/rspec/lib/spec/rails/version.rb or change the REV constant in version.rb to the current Rspec plugins Revision (1201 for the RSpec on Rails 0.7.4 plugin):

macbook-sts: $ cd vendor/plugins/rspec/lib/spec/rails
macbook-sts: $ svn pe svn:keywords version.rb
"Id Author Revision" -> "Id Author" (for example)

macbook-sts: $ svn diff -r44:45
Index: version.rb
===================================================================
--- version.rb  (revision 44)
+++ version.rb  (revision 45)
@@ -3,7 +3,7 @@
     module VERSION
       unless defined?(REV)
         # RANDOM_TOKEN: 0.644898830762263
-        REV = "$LastChangedRevision: 1201 $".match(/LastChangedRevision: (\d+)/)[1]
+        REV = "1201"
       end
     end
   end

Using Spamassassin with Rails ActionMailer

Sat Nov 25 00:00:00 +1300 2006

Running attachr.com for quite a while i noticed that the email service (you can send an email to save@attachr.com to save a code snippet) became target for spam mails.
I decided to install spamassassin to mark incoming emails for the ActionMailer service that inserts email entries into the attachr.com database.

Running a decent Debian webserver the whole task took only about 10 minutes.

First, install spamassassin

attachr:~# apt-get install spamassassin

You need to enable spamassassing in /etc/default/spamassassin.

attachr:~# vim /etc/default/spamassassin
# /etc/default/spamassassin
# Duncan Findlay

# WARNING: please read README.spamd before using.
# There may be security risks.

# Change to one to enable spamd
ENABLED=1
# ^^^^^set ENABLED to 1

Start the Spamassassin Daemon:

attachr:~# /etc/init.d/spamassassin start

To filter every incoming email with spammassasin the alias file (/etc/postfix/attachr looks like this:

cat /etc/postfix/attachr
save: "|/usr/bin/spamc|/path/to/app/script/runner 'EntryMailer.receive(STDIN.read)'"

Spamassassin marks every email with additional mail header fields (X-Spam-...). Email marked as Spam has a X-Spam-Status header field like this:

X-Spam-Status: Yes, score=7.0 required=5.0 tests=FORGED_MUA_OUTLOOK,

The Ruby code handling inbound mails has an "email" object containing the incoming email. The "x-spam-status" field contains the header value. Just match a /^Yes/ to flag the incoming mail as spam (attachr.com just ignores entries flagged as spam).
spam = true if email["x-spam-status"].to_s =~ /^Yes/

Add generated Rails files to SVN

Tue Nov 21 00:00:00 +1300 2006

./script/generate model MediaFile --svn creates your model and test files. The generated files are automatically added to your Subversion repository (after commit).

      exists  app/models/
      exists  test/unit/
      exists  test/fixtures/
      create  app/models/media_file.rb
A         app/models/media_file.rb
      create  test/unit/media_file_test.rb
A         test/unit/media_file_test.rb
      create  test/fixtures/media_files.yml
A         test/fixtures/media_files.yml
      exists  db/migrate
      create  db/migrate/002_create_media_files.rb
A         db/migrate/002_create_media_files.rb

Rails: Reload models in script/console

Fri Aug 11 00:00:00 +1200 2006

Dispatcher.reset_application! reloads your model while running script/console.

stefan@www:~$ ./script/console
Loading development environment.
>> ...
[...]
>> Dispatcher.reset_application!

Rails 1.1.5: Mandatory security patch - UDPATE

Thu Aug 10 00:00:00 +1200 2006

This is a MANDATORY upgrade for anyone not running on a very recent edge (which isn't affected by this). If you have a public Rails site, you MUST upgrade to Rails 1.1.5. The security issue is severe and you do not want to be caught unpatched. read more

Update

Rails 1.0 and prior is not affected by the latest security breach. Neither is Rails 1.1.3. read more

Update

Unfortunately another upgrade (1.1.6) is required! Do it now:

stefans@powermac:~$sudo gem install rails --source http://gems.rubyonrails.org --include-dependencies

You might want to use this vulnerability tester to test your application. Read the code carfully and try to understand what it does before you use it to test your rails application.

Be very careful to check that it wonâ€™t accidentally do something dangerous to your app.

You can find more info about this security hole here or here.

Rails-Rdoc task fails on Windows

Mon Aug 07 00:00:00 +1200 2006

Using the most recent version of Rails the rdoc rake task fails on my windows test box.

A quick and simple hack enables the rdoc task.

Replace:

Rake::RDocTask.new("appdoc") { |rdoc|
  rdoc.rdoc_dir = 'doc/app'
  rdoc.title    = "Rails Application Documentation"
  rdoc.options << '--line-numbers --inline-source'
  rdoc.rdoc_files.include('doc/README_FOR_APP')
  rdoc.rdoc_files.include('app/**/*.rb')
}

with:

task "appdoc" do
   require 'rdoc/rdoc'
   rdoc = RDoc::RDoc.new
   options = Array.new
   #options << '-a'   # parses all methods (include protected, private)
   #options << '-cUTF-8' # you may want to set the charset
   options << '-odoc/app'
   options << '--line-numbers'
   options << '--inline-source'
   options << '-Thtml'
   options << 'doc/README_FOR_APP'
   Dir.glob('app/**/*.rb').each do |file|
    options << file
   end
   rdoc.document(options)    
end

Update: see http://dev.rubyonrails.org/ticket/2018 for further information.

Capistrano: Hide SVN Files

Fri Jul 07 00:00:00 +1200 2006

Do you use Capistrano to deploy your Rails app?

Capistrano uses subversion to deploy rails applications. Unfortunately this means that the hidden subversion directories (usually .svn) are accessible via HTTP. To prevent the exposure of .svn files or directories you are strongly encouraged to add the following lines to your server configuration:

Apache:

# Don't expose .svn files/directories    
<DirectoryMatch "^/.*/\.svn/">
    Order deny,allow
    Deny from all
</DirectoryMatch>

Using a (mod_rewrite) rewrite rule:

RewriteRule \.svn/.*   -   [F]

Lighttpd:

$HTTP["url"] =~ "/\.svn/" {
    url.access-deny = ( "" )
}

Restart lighttpd and Rails

Tue May 16 00:00:00 +1200 2006

Small shell script to restart lighttpd. The script kills all dispatch.fcgi processes and restarts the webserver.

#!/bin/bash
# Autor: Stefan Saasen 

# lighttpd stop
#kill -KILL $(ps -ax | grep "lighttpd.conf" | \
                             grep -v "grep" | awk '{print $1}')
echo "Stopping lighttpd..."
sudo /etc/init.d/lighttpd stop

# Ruby stop
echo "kill dispatch.fcgi processes..."
for i in $(ps aux | egrep "^www-data(.*)dispatch.fcgi" | \
             grep -v "grep" | awk '{print $2}'); do kill -KILL $i; done

# lighttpd start
echo "start lighttpd"
sudo /etc/init.d/lighttpd start

Daedalus considered 'verboten'

Tue Apr 25 00:00:00 +1200 2006

See: http://weblog.textdrive.com/article/185/happy-thanksgiving-daedalus-is-for-dinner , http://weblog.textdrive.com/article/178/fun and http://forum.textdrive.com/viewtopic.php?id=7081 for further information.

Freeze Rails

Sat Apr 01 00:00:00 +1200 2006

As soon as you deploy a rails app on the production server you should consider to "freeze" your Rails. That means to put the Rails framework in a subdirectory (vendor/rails) of your application.

As of Rails version 0.14.x you just need to run rake freeze_gems inside your app directory to unpack the current gems into vendor/rails.

Another way to achieve this is to check out the desired version into vendor/rails using svn.

Browse http://dev.rubyonrails.org/svn/rails/tags/ and check out the desired version using

stefan@www:~$ cd /path/to/your/app/vendor
stefan@www:~$ svn co http://dev.rubyonrails.org/svn/rails/tags/rel_1-0-0/ \
rails

In case your application already is under version control using subversion (as i hope it is) you can set svn:external to fetch the framework into vendor/rails.

stefan@www:~$ cd /path/to/your/app/vendor
stefan@www:~$ svn ps svn:externals \
"rails http://dev.rubyonrails.org/svn/rails/tags/rel_1-0-0/" .

Ruby Mysql on Mac OS X (Panther)

Sun Dec 11 00:00:00 +1300 2005

Any problems installing the Ruby mysql gem?

[sts@powerbook mysql-2.7]$ sudo gem install mysql [...] ERROR: Failed to build gem native extension.

[sts@powerbook mysql-2.7]$ cd /usr/local/lib/ruby/gems/1.8/gems/mysql-2.7 [sts@powerbook mysql-2.7]$ sudo ruby extconf.rb --with-mysql-lib=/usr/local/mysql/lib
--with-mysql-include=/usr/local/mysql/include [sts@powerbook mysql-2.7]$ sudo make [sts@powerbook mysql-2.7]$ sudo make install

How to change the asset path in RubyOnRails

Sat Dec 10 00:00:00 +1300 2005

If you don't want to name your asset directories for javascript files and stylesheets "javascripts" and "stylesheets" here is how you can change the default path settings. Add the following snippet to your app/controllers/application.rb:

module ActionView
  module Helpers
    module AssetTagHelper
      # Returns path to a javascript asset. Example:
      #
      #   javascript_path "xmlhr" # => /javascripts/xmlhr.js      
      def javascript_path(source)
        compute_public_path(source, 'js', 'js')        
      end
      # Returns path to a stylesheet asset. Example:
      #
      #   stylesheet_path "style" # => /stylesheets/style.css
      def stylesheet_path(source)
        compute_public_path(source, 'css', 'css')
      end      
    end
  end
end

Logrotate auf FreeBSD: newsyslog

Wed Oct 19 00:00:00 +1300 2005

newsyslog is a highly configurable program for managing and archiving log files.

/usr/sbin/newsyslog -r -f /home/sts/etc/newsyslog.conf

Ausgabe des Outputs ohne die Logfiles zu bearbeiten:

/usr/sbin/newsyslog -r -n -f /home/sts/etc/newsyslog.conf

[1] Link: http://www.weird.com/~woods/projects/newsyslog.html
[2] Manpage newsyslog newsyslog
[3] Manpage newsyslog.conf newsyslog.conf

Rails: Setting Content Type

Tue Jun 07 00:00:00 +1200 2005

class ApplicationController &lt; ActionController::Base

  before_filter :set_default_content_type

  def set_content_type(content_type)
    @response.headers["Content-Type"] = content_type
  end

  def set_default_content_type
    set_content_type("text/html; charset=utf-8")
  end
end

Setting the charset in a Rails App

Thu May 26 00:00:00 +1200 2005

before-filter in application.rb:

before_filter :set_charset
def set_charset
    @headers["Content-Type"] = "text/html; charset=ISO-8859-15" 
end

Howto Install Complete Ruby On Debian

Wed May 18 00:00:00 +1200 2005

The following command will install all packages and libraries with "ruby1.8" in the package name, which should lead to a more or less complete installation of ruby on your machine:
apt-get install `apt-cache search -n ruby1.8 | cut -d ' ' -f 1` irb1.8 rdoc1.8 ri1.8

A few additional packages will be listed as "recommended" or "suggested" before the download/installation starts. You can cut/paste these names into the above command line; some of these packages won't be available however, so you will have to reedit the resulting command line until it is accepted by apt-get.

Ruby and Rails on Mac OS X Tiger

Sun Apr 24 00:00:00 +1200 2005

ruby -v

[powermac:~] sts% ruby -v ruby 1.8.2 (2004-12-25) [powerpc-darwin8.0]

Readline installieren curl ftp://ftp.gnu.org/gnu/readline/readline-5.0.tar.gz | tar xfz cd readline-5.0 ./configure cd ./shlib; sed -e 's/-dynamic/-dynamiclib/' Makefile > Makefile_; mv Makefile_ Makefile; cd .. make sudo make install sudo rm /usr/lib/libreadline* sudo ln -s /usr/local/lib/libreadline* /usr/lib/

lighthttp Apache and RubyOnRails

Tue Mar 22 00:00:00 +1200 2005

Using Apache as a proxy to serve a Rails on lighttpd application

<Virtualhost xxx.xxx.xxx.xxx>
        ServerAdmin s@example.com
        ServerName rails.juretta.net
        ServerAlias www.rails.juretta.net

        ProxyPass / http://localhost:81/
        ProxyPassReverse / http://localhost:81
        # disable public proxy usage
        <LocationMatch "^[^/]">
                Deny from all
        </LocationMatch>
        ProxyRequests Off
</Virtualhost>