It’s time to take the pulse of the Ruby on Rails hosting ecosystem again.
Take the 2014 Ruby on Rails hosting survey
28 Mar 2014
28 Mar 2014
It’s time to take the pulse of the Ruby on Rails hosting ecosystem again.
28 Mar 2014
An overview of how we're using Capybara Webkit to mock Chargify for a Ruby on Rails application.
27 Mar 2014
In my last article I detailed how to run PHP through a Rack server. This works fairly well until you try to sign into the Wordpress admin section.
The problem is that Wordpress stores site URLs in the database and it will use these for some redirections. Luckily, with a few Rake tasks you can painlessly override them.
26 Mar 2014
We work primarily in Ruby on Rails, but every once and a while a client will need us to fix a critical bug in an existing PHP/Wordpress app that is slated to be deprecated.
We use Pow on our development machines. This is great for developing Rails applications but it does't play so well with other stacks that require port 80 to run.
There are some instructions out on the web that show you how to use Apache in conjunction with Pow so that both apps can be served simultaneously. This seemed like a little too much overhead, so I started poking around for a solution.
Did you know that PHP comes with an embedded web server? Yup, it's built in as of PHP 5.4 and OS X Mavericks comes with it pre-installed. If you're looking for a newer version, you can install 5.5 using homebrew-php.
Now that we don't need Apache to run PHP, we'll need something to proxy requests back to our embedded server. For this, I used Rack.
I setup a simple Gemfile with the following gems:
gem 'rack'
gem 'rack-legacy'
Then it's time to create our rackup file. Create a file titled config.ru and add the following:
require 'rubygems'
require 'bundler'
Bundler.setup
require 'rack'
require 'rack-legacy'
use Rack::ShowExceptions
use Rack::Legacy::Index
use Rack::Legacy::Php
run Rack::File.new Dir.getwd
That's it! Place this file into the root of your PHP project and you should be able to call rackup
from the terminal to start the Rack server. By default this will be at http://0.0.0.0:9292.
There is one caveat though. At this time, Rack Legacy, will only respond to requests if it thinks that it's a valid PHP file.
From the Rack Legacy library this is called before a request is proxied:
def valid? path
return false unless path =~ /\.php/
path = path[1..-1] if path =~ /^\//
path = path.split('.php', 2)[0] + '.php'
path = ::File.expand_path path, @public_dir
::File.file? path
end
In our case this wasn't working for us because of the way that pretty URLs were being handled by Wordpress. In order to get around that, I monkey-patched Rack Legacy to simply pass all requests back to PHP.
Here is what I added to the config.ru file to make this happen:
class Rack::Legacy::Php
def valid? path
return true
end
end
With that change, the Wordpress site was working as expected.
25 Mar 2014
I just ran into an issue with a branch that had a tag with the same name as the branch and it was causing strange issues with branching off of that, so this was very helpful.
20 Mar 2014
It's no mystery that client side JS frameworks are here to stay, and while they do offer many benefits. The drawbacks can be daunting, especially if you have an application with any traction already.
Here Airbnb developer Spike Brehm (@spikebrehm) lays out what he sees as the future, the Isomorphic Javascript MVC...
Oh and while we are at it. Welcome to Portland @Airbnb!
4 Mar 2014
In a project weโre currently working on, we created a form builder that could publish complex forms that include features such as:
27 Feb 2014
While checking out the Wistia blog, I came across this post that piqued my interest. It resonated with me since I have been taking care of most of our remarketing campaigns lately (sorry if you are inundated with kittens now). I hate how tracking URLs look. Filled with utm_source, utm_campaign and such. Not a clean URL by any standard. However, this is what we need to use to track these campaigns. A necessary evil, I suppose.
What Fresh URL does is delete the tracking code after those tracking services have registered their information. This gives your visitors a nice clean URL if they are moved to share the page. This prevents false page tracking and just makes the internet a nicer place to be.
27 Feb 2014
Sandi Metz explained her 5 rules (and why they matter) at Future Talk at New Relic on Monday the 24th.
21 Feb 2014
Justin Weiss posted a great article on how you can add quick and easy filtering for Rails models without bloating your controllers or models beyond adding scopes here
Have a project that needs help?