Geordi 1.3 released

Changes:

  • Geordi is now (partially) tested with Cucumber. Yay!
  • geordi cucumber supports a new @solo tag. Scenarios tagged with @solo will be excluded from parallel runs, and run sequentially in a second run
  • Support for Capistrano 2 AND 3 (will deploy without :migrations on Capistrano 3)
  • Now requires a .firefox-version file to set up a test firefox. By default now uses the system Firefox/a test Chrome/whatever and doesn't print warnings any more.
  • geordi deploy --no-migrations (aliased -M): Deploy with `cap ...

How to migrate CoffeeScript files from Sprockets to Webpack(er)

If you migrate a Rails application from Sprockets to Webpack(er), you can either transpile your CoffeeScript files to JavaScript or integrate a CoffeeScript compiler to your new process. This checklist can be used to achieve the latter.

  1. If you need to continue exposing your CoffeeScript classes to the global namespace, define them on window directly:
-class @User
+class window.User
  1. Replace Sprocket's require statement with Webpacker's...

Making ZSH the default shell on Ubuntu 20.04

ZSH is an alternative command line shell that includes some features like spelling correction, cd automation, better theme, and plugin support. You can replace Bash with ZSH like following:

sudo apt-get install zsh

Setting ZSH as default login shell

sudo usermod -s /usr/bin/zsh $(whoami)

Opening a new terminal window will show you a dialog where you can configure your initial ZSH config (Option 2 recommended).

Afterwards you can install the plugin manager Oh-My-ZSH and select a prop...

Debug Ruby code

This is an awesome gadget in your toolbox, even if your test coverage is great.

  • gem install ruby-debug (Ruby 1.8) or gem install debugger (Ruby 1.9)
  • Start your server with script/server --debugger
  • Set a breakpoint by invoking debugger anywhere in your code
  • Open your application in the browser and run the code path that crosses the breakpoint
  • Once you reach the breakpoint, the page loading will seem to "hang".
  • Switch to the shell you started the server with. That shell will be running an irb session where you can step thr...

Interacting with a Microsoft Exchange server from Ruby

Microsoft Exchange service administrators can enable Exchange Web Services (EWS) which is a rather accessible XML API for interacting with Exchange. This allows you to read and send e-mails, create appointments, invite meeting attendees, track responses, manage to-do tasks, check user availability and all other sorts of things that are usually only accessible from Outlook.

You can implement an EWS by hand-rolling your XML (the [docs](http://msdn.microsoft.com/en-us/...

Custom RSpec matcher for allowed values (or assignable_values)

In contrast to RSpec's included allow_value matcher, the attached matcher will also work on associations, which makes it ideal for testing assignable_values.

Usage example

describe Unit do

  describe '#building' do
    it 'should only allow buildings that a user has access to' do
      building = build(:building)
      other_building = build(:building)
      unauthorized_building = build(:building)

      power = Power.new(build(:user))

      Power.with_power(power) do
        expect(power).to receive(:buildings).at_least...

Upgrading Capybara with deprecated Integer selectors

Capybara added a deprecation warning in version 3.35.3 (version from 2019) that shows up if your selector is not of type String or Symbol.

Example:

click_link(10) # bad
click_link("10") # good

You might encounter this error e.g. in a pagination step or similar where you want to click on numbers. To figure out where this deprecation warning comes from try to run the tests with a step output.

bundle exec parallel_cucumber --test-options "--format=pretty" feature

The deprecation message looks like following:

Locator In...

Protip: Clone large projects multiple times

Large projects usually have large test suites that can run for a long time.
This can be annoying as running tests blocks you from picking up the next story -- but it doesn't have to be that way!

Simply clone your project's repo twice (or even more often).

When your work on a feature branch is done, simply push that branch and check it out on your 2nd copy to run tests there.
You can pick up a new story and work on that on your "main" project directory.

If you do it right, you will even be able to run tests in both your 2nd copy and your m...

Mysql::Error: SAVEPOINT active_record_1 does not exist: ROLLBACK TO SAVEPOINT active_record_1 (ActiveRecord::StatementInvalid)

Possible Reason 1: parallel_tests - running more processes than features

If you run old versions of parallel_tests with more processes than you have Cucumber features, you will get errors like this in unexpected places:

This is a bug caused by multiple processes running the same features on the same database.

The bug is fixed in versions 0.6.18+.

Possib...

Case Study: Analyzing Web Font Performance

Table of contents of the linked article:

What are Web Fonts?

  • Advantages of Web Fonts
  • Disadvantages of Web Fonts
    • Fallback Fonts
    • CSS3 @font Declaration Example
    • Fallback Font Example
    • Render Blocking and Critical Rendering Path
    • FOIT

Optimizing Web Font Delivery Further

  • Prioritize Based On Browser Support
  • Choose Only Styles You Need
  • Character Sets
  • Host Fonts Locally or Prefetch
  • Store in LocalStorage with Base64 Encoding
  • Another Method

Web Font Pe...

Geordi 1.0 released

Geordi 1.0 features a command line application geordi, that holds most of Geordi's previous commands.

New features

  • command help and usage examples right within geordi (geordi help and geordi help <command>)

  • quick command access: type just the first few letters of a command, e.g. geordi rs or geordi dev[server]

  • command dependencies, e.g. geordi rspec invokes geordi bundle-install (which bundles only if needed)

  • no cluttered /usr/bin, but all commands in one handy tool

  • template for easily adding new...

Always store your Paperclip attachments in a separate folder per environment

tl;dr: Always have your attachment path start with :rails_root/storage/#{Rails.env}#{ENV['RAILS_TEST_NUMBER']}/.


The directory where you save your Paperclip attachments should not look like this:

storage/photos/1/...
storage/photos/2/...
storage/photos/3/...
storage/attachments/1/...
storage/attachments/2/...

The problem with this is that multiple environments (at least development and test) will share the same directory structure. This will cause you pain eventually. Files will get overwritten and...

Webmock normalizes arrays in urls

Typhoeus has a different way of representing array params in a get request than RestClient.

Typhoeus: http://example.com/?foo[0]=1&foo[1]=2&foo[2]=3
RestClient: http://example.com/?foo[]=1&foo[]=2&foo[]=3

Webmock normalizes this url when matching to your stubs, so it is always http://example.com/?foo[]=1&foo[]=2&foo[]=3. This might lead to green tests, but in fact crashes in real world. Rack::Utils.build_nested_query might help to build a get re...

Using Spring and parallel_tests in your Rails application

You want Spring for super-fast binstubs like bin/rails or bin/rspec which avoid Rails boot time.
You want parallel_tests to speed up full test runs of large test suites.

Unfortunately, you do not want parallel_tests to use your Spring binstubs as those parallelized tests will share data and/or loose some information. There are some issues about this on GitHub and there is a suggested [workaround](https:...

Rails: How to check if a certain validation failed

If validations failed for a record, and you want to find out if a specific validation failed, you can leverage ActiveModel's error objects.
You rarely need this in application code (you usually just want to print error messages), but it can be useful when writing tests.

As an example, consider the following model which uses two validations on the email attribute.

class User < ApplicationRecord
  validates :email, presence: true, uniqueness: true
end

Accessing errors

Let's assume we have a blank user:

user = Us...

How to bind an event listener only once with Unpoly

You can use Unpoly's up.on with a named listener function and immediately unbind this event listener with { once: true }:

up.on('up:fragment:inserted', { once: true }, function () { ... })

In Unpoly 1 you can immediately unregister the listener with up.off:

up.on('up:fragment:inserted', function fragmentInsertedCallback() {
  up.off('up:fragment:inserted', fragmentInsertedCallback)
  // ... code for the callback function, which should run only once
})

Exam...

Ruby 1.9 or Ruby 2.0 do not allow using shortcut blocks for private methods

Consider this class:

class Foo

  private
  
  def test
    puts "Hello"
  end
  
end

While you can say create a block to call that method (using ampersand and colon) on Ruby 1.8, ...

1.8.7 > Foo.new.tap(&:test)
Hello
=> #<Foo:0x1e253c8> 

... you cannot do that on Ruby 1.9 or 2.0:

1.9.3 > Foo.new.tap(&:test)
NoMethodError: private method `test' called for #<Foo:0x00000001e8c258>

^
2.0.0 > Foo.new.tap(&:test)
NoMethodError: private method `test' called for #<Foo:0x000000027bc738...

An incomplete guide to migrate a Rails application from paperclip to carrierwave

In this example we assume that not only the storage gem changes but also the file structure on disc.

A general approach

Part A: Create a commit which includes a script that allows you to copy the existing file to the new file structure.

Part B: Create a commit which removes all paperclip logic and replace it with the same code you used in the first commit

Part A

Here are some implementation details you might want to reuse:

  • Use the existing models to read the files from
  • Use your own carrierwave models to write t...

GitHub Actions: Retrying a failing step

If you have a flaky command you can use the nick-invision/retry to re-try a failing command, optionally with a timeout:

---
...
jobs:
  test:
    ...
    steps:
    - name: Run tests
      uses: nick-invision/retry@v2
      with:
        timeout_seconds: 30
        max_attempts: 3
        command: bundle exec rake spec

How to fix gsub on SafeBuffer objects

If you have an html_safe string, you won't be able to call gsub with a block and match reference variables like $1. They will be nil inside the block where you define replacements (as you already know).

This issue applies to both Rails 2 (with rails_xss) as well as Rails 3 applications.

Here is a fix to SafeBuffer#gsub. Note that it will only fix the $1 behavior, not give you a safe string in the end (see below).

Example

def test(input)...

The new Modularity 2 syntax

We have released Modularity 2. It has many incompatible changes. See below for a script to migrate your applications automatically.

There is no does method anymore

We now use traits with the vanilla include method:

class Article < ActiveRecord::Base
  include DoesTrashable
end

When your trait has parameters, use square brackets:

class Article < ActiveRecord::Base
  include DoesStripFields[:name, :brand]
end

Note how you ...

Upgrading Ruby from 1.8.7 to 2.3.5

Suggested Workflow

Set the ruby version in .ruby-version to 2.3.5, then perform these steps one by one, fixing errors as they occur:

  1. Update gems as listed below, and bundle
  2. Boot a Rails console - see below for a list of changes you will probably need
  3. Run Specs with --backtrace option
  4. Run Cucumber features (with Geordi's --debug option)
  5. When all tests are green, look through your Gemfile and remove as many version constraints as possible.
  6. Boot the application in different environements to spot further issues, e...

parallel_tests: Disable parallel run for tagged scenarios

Note: This technique is confusing and slows down your test suite.


Copy the attached code to features/support. This gets you a new Cucumber tag @no_parallel which ensures that the tagged scenario does not run in parallel with other scenarios that are tagged with @no_parallel. Other scenarios not tagged will @no_parallel can still run in parallel with the tagged test. Please read the previous sentence again.

This can help when multiple test processes that access a single resource that is hard to shar...

Jasmine: Mocking API requests in an Angular service spec

Situation: You want to write a spec for a function inside an Angular service. This function at some point makes an API request and acts upon response. Notably, your Angular app employs uiRouter, although it is not used nor actually required for this test.

Working test setup

# Capitalized expressions are intended to be replaced with YOUR values

describe 'SERVICE', ->

  beforeEach ->
    module 'MODULE'

    module ($urlRouterProvider) ->
      # Prevent uiRouter's initialization, i.e. do not sync the current URL
      # with its r...