cavewall.jaguardesignstudio.comThe Cave Wall - Jaguar Design Studio Engineering Blog

Running a Ruby Executable Outside of Bundler May 9th, 2014 - Brendon Rapp One of our internal use gems includes some executable scripts which are basically wrappers around other gem executables. Most of these gems are part of the application’s Gemfile, but sometimes, they are globally installed gems. An example of one such case is Mailcatcher, which, due to the gem’s own dependencies (mainly ActiveSupport), should not be included in a Rails app’s Gemfile . The problem we ran into is that, when trying to execute the global Mailcatcher from this wrapper script, it would actually try to execute within the Bundler environment, and resulted in this error: mailcatcher is not part of the bundle. Add it to Gemfile. The script looked something like this: #!/usr/bin/env ruby ... some setup stuff ... exec 'mailcatcher -f -v' What we needed to do was break out of the Bundler environment with Bundler.with_clean_env . The Bundler documentation describes as such: Any Ruby code that opens a subshell (like system , backticks, or %x{} ) will automatically use the current Bundler environment. If you need to shell out to a Ruby command that is not part of your current bundle, use the with_clean_env method with a block. Updating our wrapper script as such solved the issue: #!/usr/bin/env ruby ... some setup stuff ... Bundler.with_clean_env { exec "mailcatcher -f -v" } ActiveAdmin Filters with Ransack May 1st, 2014 - russ Recently, in the past three months or so, ActiveAdmin has updated their master branch to add Rails 4.1 support and migrate away from MetaSearch and on to Ransack for their filters. If you use the stock filters or shallow relation filters then more than likely your current ActiveAdmin filters will continue to work. However, if you need more complex filtering or previously had defined “search_methods” in your models, you will need to update your code to make use of Ransack’s “ransacker” method instead. Your filter declarations in your ActiveAdmin models are the same format and can remain untouched but you may have to update the filter names and labels. Like before, you will require changes in two files for a complex filter: one in your ActiveAdmin register and one in the associated model. Below is an example where an Order exists with many Products which each have one ProductColor and we want a drop-down select filter at the Order level for ProductColor: /app/admin/order.rb ActiveAdmin.register Order do ... filter :product_color_in, :as => :select, :label => 'Containing Product Color', :collection => proc { ProductColors.order(:name) } ... end /app/models/order.rb class Order < ActiveRecord::Base has_many :products # and Product belongs_to a ProductColor ... ransacker :product_color, formatter: proc { |selected_pc_id| results = Order.has_pc(selected_pc_id).map(&:id) results = results.present? ? results : nil }, splat_params: true do |parent| parent.table[:id] end def self.has_pc(pc_id) self.joins(:products).where(products: {product_color_id: pc_id}) end ... end Sources https://github.com/gregbell/active_admin , specifically the top “1.0.0” section. https://github.com/activerecord-hackery/ransack/issues/345 https://github.com/activerecord-hackery/ransack/issues/36 (3 years old but it helped in a solution) http://nikhgupta.com/code/activeadmin/custom-filters-using-ransacker-in-activeadmin-interfaces/ Cloud9 with a Private Server September 13th, 2013 - russ As far as developers go, I’m not too picky about the tools I use. I find this allows me to try out the latest and greatest without too much attachment to a specific toolkit. One type of toolkit that has been gaining popularity is a cloud-based IDE. I’ve always been interested in a centralized workspace since I tend to blow through devices and workstations (usually to try out some new Linux distribution) and frequently have to reconfigure for the new environments. Yes, I could have just used dual-boot or swap hard drives but I really don’t care that much. All my important data is already in the cloud, through services like Google Drive, Dropbox, and Bitbucket. In a way, I’m already set up for cloud-based solutions and for most people that may be the most difficult part, especially if you have a lot of data. If that’s the case, not all is lost. You can use many of these IDEs with any common git repository or in some cases, with SSH access to your own private server. Cloud9 is one of these cloud-based IDEs and they provide a private PaaS (Platform as a Service) instance, or you can make use of your own FTP/SSH server. They also support directly deploying to Heroku, Azure, and Cloud Foundry. In my case, I use my own PaaS instance and connect to it through SSH. This does have one prerequisite: NodeJS. In order for Cloud9 to access your workstation, you will need to not only configure your SSH server and keys, but you will also need to set up a NodeJS server (along with any firewall rules). Cloud9 provides excellent documentation . Once these are configured, Cloud9 provides a full-featured IDE, including a decent terminal (though it’s not as responsive as other web-based terminals) that allows you to develop in any language you want. This setup has been working well for me for the past few weeks across multiple platforms: Linux, OS X, Windows, and even my Chromebook. This is especially useful for quick fixes and testing when on the road with my $210 Chromebook, instead of taking my $1,200 MacBook Air. Taking advantage of the recent SaaS (Software as a Service) and modern web browser improvements can save money in equipment, maintenance, security, and training. I look forward to seeing it improve even more. Cave Lunch – January 2013 March 7th, 2013 - Brendon Rapp Starting in January, the engineering team instated monthly “tech lunch” meetings. On the final Friday of each month, the team will meet for lunch, and each member will present a 5-10 minute “lightning talk” on some development topic of interest. Below are the slides from January’s talks: Selenium IDE by Renee Wall POWering Through Rails/Rack Application Development by Jeremy Ebler Why Zsh is Cooler than Your Shell by Brendon Rapp 5 Tricks App-Makers Use to Boost In-App Purchases by Zach Archer The Normal Computer User’s Guide to Safe Computing June 15th, 2012 - Brendon Rapp Your CHILDREN may be at risk! Find out tonight at 11:00. In the past week, LinkedIn, eHarmony, and Last.fm were each subject to a serious security breach , resulting in the leakage of many user passwords. Not coincidentally, in the same week, Facebook started pushing a security tips page to all of its users. The cold, hard truth is that it’s more likely than not that you will have some Internet account that you use end up having its password leaked out at some point in time. Whether or not that results in a minor temporary inconvenience, or more dramatic identity theft, is up to you. Unfortunately, there has not been a very good effort to teach everyday computer users how to keep themselves safe. Equally unfortunate, too many everyday computer users ignore advice when they are taught, and fail to recognize their own complicity when they do become victims. It does not take a great deal of effort to dramatically increase your digital security. A few small habit adjustments and some tools to help are all it takes. The following is a brief, opinionated guide for “normal, everyday” computer users. People of a higher technical background will likely have their own preferences, and that’s fine, but this is a guide that we consider to be ideal for the average user. 1. Practice Good Password Security The rules for passwords: No password should ever be used twice Passwords should be long and random Don’t try and memorize your passwords – use a password manager People often use simple passwords, and use them over and over, because it’s difficult to remember a whole bunch of random passwords. That is why password managers were invented. Instead of remembering passwords yourself,...