Rails: Do not load frameworks you don't need
Rails is split into a large number of (sub-) frameworks.
The most important and central of those are
- activesupport (extends the Ruby standard library)
- activerecord / activemodel (ORM for Rails)
- actionview / actionpack (controller / views)
- actionmailer (sends mails)
However, there are also some more situational frameworks included, such as
- actioncable (real time communications using websockets)
- actionmailbox (receives mails)
- actiontext (support for WYSIWYG text editor)
- activejob (background jobs)
- activestorage (file uplo…
Cucumber pitfall: "Around" does not apply to your "Background" steps
Around will not happen until after a feature's Background has been processed. Use Before and After to avoid that.
Details
Consider this Cucumber feature file:
Feature: Something that needs to be tested
Background:
Given a user
And I sign in
Scenario: Sign out
When I sign out
Then I should see "Signed out"
Scenario: Something else
# ...
Now, assume you have these step definitions:
Around do
puts "** Around: before yield" ...
Casting ActiveRecord scopes or instances to ActiveType extended model classes
When working with ActiveType you will often find it useful to cast an ActiveRecord instance to its extended ActiveType::Record variant.
Starting with active_type 0.4.0 you can use ActiveType.cast for this:
class User < ActiveRecord::Base
...
end
class SignUp < ActiveType::Record[User]
...
end
user = User.find(1)
sign_up = ActiveType.cast(user, SignUp)
sign_up.is_a?(SignUp) # => true
This is basically like [ActiveRecord#becomes](http://apidock.com/rails/v4.2.1/ActiveRecord/Pers…
Linux: Running a single unique instance of command + arguments
run-one is a wrapper script that won't run the given command while another instance of it is running. Is brings several utility commands that offer similar behavior.
```
NAME
run-one - run just one instance at a time of some command and unique set of arguments (useful for cronjobs, eg)
SYNOPSIS
run-one COMMAND [ARGS]
run-this-one COMMAND [ARGS]
run-one-constantly COMMAND [ARGS]
keep-one-running COMMAND [ARGS]
run-one-until-success COMMAND [ARGS]
run-one-until-failure COMMAND [ARGS]
…
Controlling how your website appears on social media feeds
When a user shares your content, a snippet with title, image, link and description appears in her timeline. By default social networks will use the window title, the first image, the current URL and some random text snippet for this purpose. This is often not what you want.
Luckily Facebook, Twitter, etc. lets you control how your content appears in the activity streams. They even have agreed on a common format to do this: OpenGraph <meta> tags that go into your HTML's <head>:
<meta property="og:url" content="http://start.makan...
Dealing with I18n::InvalidPluralizationData errors
When localizing model attributes via I18n you may run into errors like this:
I18n::InvalidPluralizationData: translation data { ... } can not be used with :count => 1. key 'one' is missing.
They seem to appear out of the blue and the error message is more confusing than helpful.
TL;DR A model (e.g. Post) is lacking an attribute (e.g. thread) translation.
Fix it by adding a translation for that model's attribute (attributes.post.thread). The error message reveals the (wrongly) located I18n data (from attributes.thread)….
Regex: Be careful when trying to match the start and/or end of a text
Ruby has two different ways to match the start and the end of a text:
^(Start of line) and$(End of line)\A(Start of string) and\z(End of string)
Most often you want to use \A and \z.
Here is a short example in which we want to validate the content type of a file attachment. Normally we would not expect content_type_1 to be a valid content type with the used regular expression image\/(jpeg|png). But as ^ and $ will match lines, it matches both content_type_1 and content_type_2. Using \A and \z will wo…
Write custom RSpec matchers
There are three ways to define your own RSpec matchers, with increasing complexibility and options:
1) Use RSpec::Matchers.define
RSpec::Matchers.define :be_a_multiple_of do |expected|
match do |actual|
actual % expected == 0
end
# optional
failure_message do |actual|
"expected that #{actual} would be a multiple of #{expected}"
end
# optional
failure_message_when_negated do |actual|
"expected that #{actual} would not be a multiple of #{expected}"
end
end
This is automat…
Matching unicode characters in a Ruby (1.9+) regexp
On Ruby 1.9+, standard ruby character classes like \w, \d will only match 7-Bit ASCII characters:
"foo" =~ /\w+/ # matches "foo"
"füü" =~ /\w+/ # matches "f", ü is not 7-Bit ASCII
There is a collection of character classes that will match unicode characters. From the documentation:
/[[:alnum:]]/Alphabetic and numeric character/[[:alpha:]]/Alphabetic character/[[:blank:]]/Space or tab/[[:cntrl:]]/Control character/[[:digit:]]/Digit/[[:graph:]]/Non-blank character (excludes spaces, co…
RubyMine users: you should be using bookmarks
RubyMine allows bookmarking lines of code. This is super-helpful when working on a complex problem.
I've been using this feature for a few years now, and so should you! :)
Here are the default Linux/Windows keystrokes. See the documentation for other keybindings.
Add an anonymous bookmark
F11
A gray checkmark will be shown in the gutter on the left.
If you press F11 again on a bookmarked line, the bookmark will be removed.
Add a named bookmark ("mnemonic")
Ctrl …
When upgrading/downgrading RubyGems and Bundler on a server, you must clear bundled gems
On application servers, gems are usually bundled into the project directory, at a location shared across deployments.
This is usually shared/bundle inside your project's root directory, e.g. /var/www/your-project/shared/bundle/.
If you can't find that, take a look at current/.bundle/config and look for BUNDLE_PATH.
When you are changing the version of RubyGems or Bundler on a system where gems are installed this way, you must wipe that bundle directory in addition to the user and system gems or gems that are already ins…
Be very careful with 301 redirects
Browsers support two kinds of redirects:
301 moved permanently302 moved temporarily
Be very careful with the 301 type. Most browsers seem to cache these redirects forever, unless you set different Cache-Control headers. If you don't have any cache control headers, you can never change them without forcing users to empty their cache.
The only fix is to keep redirecting the user to the correct page, so if you had
/page-1 --301--> /page-2
but you want
/page-1 --301-> /page-3
your only fix is to change it to
…
Ruby: How to use global variables for a conditional debugger
You can share a state in Ruby with global variables. Even if you should avoid them whenever possible, for debugging an application this could be temporary quite handy.
Example:
class User
after_save { byebug if $debug; nil }
def lock
self.locked = true
save
end
end
```
Rspec.describe User do
let(:user) { create(:user) }
before do
# Many users are created and saved in this hook, but we don't want the debugger to stop for them
end
…