ARel and ActiveRecord

My slides on ARel and ActiveRecord

Very colorful vim

It seems that xterms and screen are slowly clawing their way into the 21st century... We can now view our code in more than 16 colors. There are a few hoops you'll need to jump through to get it working.

• Make sure tmux is 256 ready, editing your ~/.tmux.conf
  
  
  
  • set -g default-terminal "screen-256color"
  
  



• Make sure screen is 256 ready, editing your ~/.screenrc
  
  
  
  • 
    

    # terminfo and termcap for nice 256 color terminal
    # allow bold colors - necessary for some reason
    attrcolor b ".I"
    # tell screen how to set colors. AB = background, AF=foreground
    termcapinfo xterm 'Co#256:AB=\E[48;5;%dm:AF=\E[38;5;%dm'
    # erase background with current bg color
    defbce "on"

  
  


• Check your terminal supports 256 colors
  
  
  
  • colortest
  
  
  • Run that in and out of screen
  
  
  • If you find that the gradient doesn't print what you expect, you may need to recompile screen with the configure option --enable-colors256
  
  
  • If that doesn't work, you may need to recompile X...
  
  


• Update your .vimrc
  
  
  
  • syntax on

set t_Co=256
  
  

You can now begin using some super fancy color schemes with vim. While in vim, use :colorscheme to test drive the many and varied options. Fortunately, colorscheme supports tab completion and it will search your system (/colors) and user (~/.vim/colors) runtime directories. Once you've settled on a scheme you like, add the command to your .vimrc.

Very nice (am I crazy, or are a bunch of these fifteen year old emacs color schemes)

• Inkpot


• Jellybeans


• Jellyx


• Herald


• Lucius

Not too shabby

• dark-ruby


• desert256


• molokai

Migrating to MySQL

Diaspora has chosen to move from NoSQL to SQL (too much relational data). As of this writing, they are only known to support MySQL, since that's what they're running on joindiaspora.com, AFAIK. Here are my hoops, on Ubuntu...
https://github.com/diaspora/diaspora/wiki/Migrating-from-MongoDB-to-MySQL

1. Update your source
   
   • git fetch upstream
   
   
   • git checkout master
   
   
   • git pull upstream master


2. Install mysql packages
   
   • apt-get install mysql-server mysql-client libmysqlclient16-dev


3. Make sure you're utf8 all the way, Diaspora is
   
   • sudo vi /etc/mysql/my.cnf
     
     1. Update the server section [mysqld]
        
        • collation_server=utf8_bin
          
          character_set_server=utf8
     
     
     2. Update the client section [mysql]
        
        • default-character-set = utf8


4. Restart MySQL
   
   • service mysql restart


5. Create your database.yml
   
   • cp config/database.yml.example config/database.yml


6. Edit your new yaml file to suit your environment
   
   • Note As of this writing, only the MySQL user root can migrate from mongo to MySQL


7. Create the necessary databases and tables
   
   • rake db:drop:all
     
     • Make sure test, development, and production all have valid credentials (for root mysql user)
     
     
     • You can safely ignore messages like:
       Couldn't drop diaspora_production : <Mysql2::Error: Unknown database 'diaspora_production'>
   
   
   • rake db:create
   
   
   • rake db:migrate
   
   
   • rake migrations:migrate_to_mysql

Voila, you should now have everything you need to help the community continue to test this massive change...

A rake task to syntax check erb, haml, rb, sass, and yaml files

Diaspora is pretty fluid right now. This means we are have some green tests, some missing tests, and other tests that check intent (not implementation). In an ideal world, I suppose test cases would cover all of our bases...

Until then, I've added a new task to my fork, check_syntax:all. This breaks down further to the subtasks check_syntax:erb, check_syntax:haml, check_syntax:haml_ruby, check_syntax:sass, and check_syntax:yaml. Most of these tasks are pretty straight forward, search the tree for .suffix files and check their syntax. The subtask check_syntax:haml_ruby uses HAML::Engine.new().precompiled (similar output to haml --debug) to generate the ruby haml generates. The resulting ruby is syntax checked... This is very important, because haml --check does not check ruby syntax.

The following is based on Hungry Machine's post

code is also available from here

namespace :check_syntax do
  require 'erb'
  require 'open3'
  require 'yaml'

  desc "Check syntax of various file types"
  task :all => [:ruby, :erb, :yaml, :haml, :haml_ruby, :sass]

  #desc 'Check syntax of .erb and .rhtml files'
  task :erb do
    (Dir["**/*.erb"] + Dir["**/*.rhtml"]).each do |file|
      next if skip_path?(file)

      Open3.popen3('ruby -Ku -c') do |stdin, stdout, stderr|
        stdin.puts(ERB.new(File.read(file), nil, '-').src)
        stdin.close
        if error = ((stderr.readline rescue false))
          puts red_string(file) + error[1..-1].sub(/^[^:]*:\d+: /, '')
        end
        stdout.close rescue false
        stderr.close rescue false
      end
    end
  end

  #desc 'Check syntax of .haml files'
  task :haml do
    haml = bin_path "haml"
    Dir['**/*.haml'].each do |file|
      next if skip_path?(file)
      next if file.match("vendor/plugins/.*/generators/.*/templates")

      Open3.popen3("#{haml} -c #{file}") do |stdin, stdout, stderr|
        if error = ((stderr.readline rescue false))
          puts red_string(file) + error
        end
        stdin.close rescue false
        stdout.close rescue false
        stderr.close rescue false
      end
    end
  end

  #desc 'Check ruby syntax of .haml files'
  task :haml_ruby do
    Dir['**/*.haml'].each do |file|
      next if skip_path?(file)
      next if file.match("vendor/plugins/.*/generators/.*/templates")

      Open3.popen3("ruby -Ku -c") do |stdin, stdout, stderr|
        stdin.puts(Haml::Engine.new(File.read(file)).precompiled)
        stdin.close
        if error = ((stderr.readline rescue false))
          puts red_string(file) + error.sub(/^[^:]*:\d+: /, '')
        end
        stdout.close rescue false
        stderr.close rescue false
      end
    end
  end

  #desc 'Check syntax of .rb files'
  task :ruby do
    Dir['**/*.rb'].each do |file|
      next if skip_path?(file)
      next if file.match("vendor/plugins/.*/generators/.*/templates")

      Open3.popen3("ruby -Ku -c #{file}") do |stdin, stdout, stderr|
        if error = ((stderr.readline rescue false))
          puts error
        end
        stdin.close rescue false
        stdout.close rescue false
        stderr.close rescue false
      end
    end
  end

  #desc 'Check syntax of .sass files'
  task :sass do
    sass = bin_path "haml", "sass"
    Dir['**/*.sass'].each do |file|
      next if skip_path?(file)
      next if file.match("vendor/plugins/.*/generators/.*/templates")

      Open3.popen3("#{sass} -c #{file}") do |stdin, stdout, stderr|
        if error = ((stderr.readline rescue false))
          puts error
        end
        stdin.close rescue false
        stdout.close rescue false
        stderr.close rescue false
      end
    end
  end

  #desc 'Check syntax of .yml files'
  task :yaml do
    Dir['**/*.yml'].each do |file|
      next if skip_path?(file)

      begin
        YAML.load_file(file)
      rescue => e
        puts red_string(file) + "#{(e.message.match(/on line (\d+)/)[1] + ':') rescue nil} #{e.message}"
      end
    end
  end

  def bin_path gem_name, bin_name=nil
    bin_name=gem_name if bin_name.nil?
    [`bundle show #{gem_name}`.chomp!, 'bin', bin_name].join("/")
  end

  def skip_path?(file)
    ["tmp", "vendor/gems", "vendor/rails"].each do |dir|
      return true if file.match(dir)
    end
    return false
  end

  def red_string(str)
    "\e[31m" + str + "\e[0m: "
  end
end

Running Diaspora, without ./script/server on Ubuntu

I wanted to free my terminal from the plethora of Diaspora* foreground daemons... I also wanted to make sure that these alpha daemons could be restarted, once they inevitably crashed. I tried using an older version of an Ubuntu init script.

The original version followed older init practices, and avoided upstart altogether. This allowed me to skip script/server, but didn't restart services. As such, I decided to yank the core team's daemontools implementation. Core functionality was there, using this approach, but websocket didn't quite work (no pushes to clients) and I noticed numerous complaints, running ps axuw|grep readproc. No amount of fiddling yielded usable/stable results...

Finally, I decided to change my service names, prefixing each name with a number, similar to init and then instituting requirements. It is redundant, but the numbering was a convenient mnemonic ... The order of execution mimics the order script/server.

Install instructions, a version of my work, and a handy script for controlling daemontools services can be found at github.

I completed work on this and two days later (one day later?) leamas completed a wholesale rewrite of the Ubuntu package installer. It seems to have tighter integration with Ubuntu's edicts and expectations...