Tailoring Rails error messages for models and attributes
Rails has generic error messages you can define in your locale .yml files. You may override those application-wide error messages using model or attribute scope like this:
en:
activerecord:
errors:
messages:
invalid: is invalid # used for any invalid attribute in the application
models:
car:
invalid: does not work # used for invalid car attributes
attributes:
driver:
invalid: not allowed to drive # used if the car's ...
Merging two arbitrary ActiveRecord scopes
Rails 3+ allows you to join two scopes from arbitrary sources by calling the merge method:
scope1 = User.where(:email => 'foo@bar.com')
scope2 = User.where(:first_name => 'hans')
merged_scope = scope1.merge(scope2)
merged_scope.to_a will now trigger a query for the combined scope chain:
SELECT `users`.* FROM `users` WHERE `users`.`email` = 'foo@bar.com' AND `users`.`first_name` = 'hans'
If you are joining two models, you can also [merge scopes for different models](http://blog.thefrontiergroup.com.au/2011/03/composing…
ActiveRecord scopes must be loaded before using RSpec's "=~" matcher
To test whether two arrays have the same elements regardless of order, you can use the =~ matcher in RSpec < 2.11:
actual_array.should =~ expected_array
If either side is an ActiveRecord scope rather than an array, you should call to_a on it first, since =~ does not play nice with scopes:
actual_scope.to_a.should =~ expected_scope.to_a
If you use RSpec >= 2.11 we recommend to use the match_array or eq matcher. There are also use cases where the contain_exactly matcher fits better.
Auto-generating plain-text bodies for HTML e-mails in Rails apps
When building an application that sends e-mails to users, you want to avoid those e-mails from being classified as spam. Most obvious scoring issues will not be relevant to you because you are not a spammer.
However, your application must do one thing by itself: When sending HTML e-mails, you should include a plain-text body or tools like SpamAssassin will apply a significant score penalty. Here is how to do that automatically.
- Add
premailer-railsto yourGemfileandbundle. - Done! …
Does <html> or <body> scroll the page?
Scrolling overflowing elements with JavaScript
HTML elements with overflow-y: auto or overflow-y: scroll will get a scrollbar when their content is higher than their own height.
When you scroll an element , the element's scrollTop property is updated with the scrollbar's new position. You can also set element.scrollTop = 30 to scroll the element to a vertical pixel position counted from the top.
Scrolling the main viewport with JavaScript
The browser's main document viewport is also scrollable by default. The element that …
Using ES6 without a transpiler
If your app does not need to support IE11, you can use most ES6 features without a build step. Just deliver your plain JavaScript without transpilation through Babel or TypeScript, and modern browsers will run them natively.
Features supported by all modern browsers include:
- fat arrow functions (
() => { expr }) -
let/const class-
async/await - Promises
- Generators
- Symbols
- Rest arguments (
...args) - Destructuring
You won't be able to use import and export, or use npm modules.
See this [ES6 compatibility mat…
Custom error messages in RSpec or Cucumber steps
Sometimes you have a test expectation but actually want a better error message in case of a failure. Here is how to do that.
Background
Consider this test:
User.last.should be_present
In case of an error, it will fail with a not-so-helpful error message:
expected present? to return true, got false (Spec::Expectations::ExpectationNotMetError)
Solution
That can be fixed easily. RSpec expectations allow you to pass an error message like this:
User.last.should be_present, 'Could not find a user!'
Now your test will…
How to emulate simple classes with plain JavaScript
If you want a class-like construct in JavaScript, you can use the module pattern below. The module pattern gives you basic class concepts like a constructor, private state, public methods.
Since the module pattern only uses basic JavaScript, your code will run in any browser. You don't need CoffeeScript or an ES6 transpiler like Babel.
A cosmetic benefit is that the module pattern works without the use of this or prototypes.
Example
Here is an example for a Ruby class that we want to translate into Javascript using the module patter…
Canceling promises
The cancelable promises proposal was withdrawn some time ago.
The new standard way is that your long-running function take a AbortSignal { signal } property. The caller can use this signal to send an abort request to your function. Upon receiving the request, your function should reject its promise with an error.
Async browser functions like fetch() reject their promises with a `new DOMException('Message here', 'Abo…
Summarizing heredoc in Ruby and Rails
This card tries to summarize by example the different uses of heredoc.
- In Ruby
<<vs.<<-vs.<<~ - In Rails
strip_heredocvs.squish
strip_heredoc should be used for a text, where you want to preserve newlines. (multi-line -> multi-line)
squish should be used for a text, where you want to squish newlines. (multi-line -> one-line)
Ruby 2.3+
def foo
bar = <<~TEXT
line1
line2
line3
TEXT
puts bar.inspect
end
foo => "line1\nline2\nline3\n"
Read more: [Unindent HEREDOCs in Ruby 2.3](/makan…
How to overwrite and reset constants within Cucumber features
In order to save the original value of a constant, set the new value and restore the old value after a scenario was completed, you can use the following helper. It takes care of saving the old constant value, setting the new one without throwing warnings and resets the value with an After hook.
This module also enables you to introduce new global constants.
Since these newly defined constants do not have any value to be reset to,
they simply are deleted (remove_const) once the respective Cucumber step finishes.
You can copy the file at…
Gatekeeping: Guide for developer
Note: This process is tailored to our specific needs and tools at makandra. While it will certainly not apply to all (especially larger teams), we think it is a helpful starting point. Compare also the Gatekeeping: Guide for gatekeeper card.
In order to reduce the number of rejects we get from clients, we want to review all code written before it goes to the staging server.
If your project manager wants to do gatekeeping o…
Capistrano: Speeding up Sprockets' asset compile during deploy
Remember How to skip Sprockets asset compile during Capistrano deployment and Automatically skipping asset compilation when assets have not changed? Turns out there is an even better way to speed up Capistrano deployments with asset compilation – and it's even simpler.
Adding the Sprockets cache directory to shared directories
Sprockets is the library behind Rails' asset pipeline. It keeps a cache of already compiled files, so when a single file changes, it only needs to rebuild that one. This is …
What collapsing margins are, how they work and when margins do not collapse
What collapsing margins are
Imagine you have 2 HTML boxes. The first one has a margin-bottom of let's say 30px and the second one a margin-top of 20px. After rules of collapsing margins have been applied we have a margin of 30px (not 50px) between these two boxes (see this article for an illustration) . This is because no addition of both margins takes place but the maximu…
Rails index route for resources named after uncountable substantives
Using uncountable resources is not recommended as it breaks Rails' magic, e.g. when using form_for. You'll always be better off using simple pluralizable resources.
Rails automatically creates path names for routes defined via the resource method. When you put resource 'user' into config/routes.rb, you can call users_path and get the path to the index action in the UsersController: /users.
However, if you have an uncountable resource like Sheep, you cannot access the index action via sheep_path, because it will…