In this first article of a new series on the Asset Pipeline in Rails, I’d like to discuss a few high-level concepts that are handy to fully grasp what the Asset Pipeline has to offer and what it does under the hood. The main things it manages for you are concatenation, minification, and preprocessing of assets. As a beginner, you need to familiarize yourself with these concepts as early as possible.
Topics
- Asset Pipeline?
- Organization
- Concatenation & Compression?
- High-Level Languages
Asset Pipeline?
The Asset Pipeline is not exactly news to people in the business, but for beginners it can be a little tricky to figure out quickly. Developers don’t exactly spend a ton of time on front-end stuff, especially when they start out. They are busy with lots of moving parts that have nothing to do with HTML, CSS, or the often bashed JS bits.
The Asset Pipeline can help by managing the concatenation, minification, and preprocessing of assets—for example, assets that are written in high-level languages like CoffeeScript or Sass.
This is an important step in the build process for Rails apps. The Asset Pipeline can give you a boost not only in quality and speed but also in convenience. Having to set up your own tailored build tool might give you a few more options for fine-tuning your needs, but it comes with an overhead that might be time-consuming for beginners—maybe even intimidating. In this regard, we could talk about the Asset Pipeline as a solution that fosters convenience over configuration.
The other thing that should not be underestimated is organization. The pipeline offers a solid framework to place your assets. On smaller projects this might not seem that important, sure, but bigger projects might not recover easily from going in the wrong direction on subjects like this. This might not be the most common example in the world, but just imagine a project like Facebook or Twitter having bad organization for their CSS and JS assets. Not that hard to imagine that this would breed trouble and that people who have to work and build on such a basis would not have an easy time loving their jobs.
As with many things in Rails, there is a conventional approach of how to handle assets. This makes it easier to onboard new people and to avoid having developers and designers being too obsessed with bringing their own dough to the party.
When you join a new team, you want to be able to get up to speed quickly and hit the ground running. Needing to figure out the special sauce of how other people organized their assets is not exactly helpful with that. In extreme cases, this can even be regarded as wasteful and can burn money unnecessarily. But let’s not get too dramatic here.
So this article is for the beginners among you, and I recommend taking a look at the pipeline right away and getting a good grip on this topic, even if you don’t expect to work on markup or front-endy things much in the future. It’s an essential aspect in Rails and not that huge of a deal. Plus your team members, especially designers who spend half their lives in these directories, will put less funny stuff in your coffee if you have some common ground that does not conflict with each other.
Organization
You have three options for placing your assets. These divisions are more of a logical kind. They don’t represent any technical restrictions or limitations. You can place assets in any of them and see the same results. But for smarter organization, you should have a good reason to place content in its proper place.
app/assets
app/assets/
├── images
├── javascripts
│ └── application.js
└── stylesheets
└── application.css
The app/assets folder is for assets that are specifically for this app—images, JS and CSS files that are tailor made for a particular project.
lib/assets
lib/assets/
The lib/assets folder, on the other hand, is the home for your own code that you can reuse from project to project—things that you yourself might have extracted and want to take from one project to the next—specifically, assets that you can share between applications.
vendor/assets
vendor/assets/ ├── javascripts └── stylesheets
vendor/assets is another step outward. It’s the home for assets that you reused from other, external sources—plugins and frameworks from third parties.
These are the three locations in which Sprockets will look for assets. Within the above directories, you are expected to place your assets within the /images, /stylesheets and /javascripts folders. But this is more of a conventional thing; any assets within the */assets folders will be traversed. You can also add subdirectories as needed. Rails won’t complain and precompile their files anyway.
There is an easy way to take a look at the search path of the asset pipeline. When you fire up a Rails console with rails console, you can inspect it with the following command:
Terminal
Rails.application.config.assets.paths
What you will get returned is an array with all the directories of assets that are available and found by Sprockets—aka the search path.
=> ["/Users/example-user/projects/test-app/app/assets/images", "/Users/example-user/projects/test-app/app/assets/javascripts", "/Users/example-user/projects/test-app/app/assets/stylesheets", "/Users/example-user/projects/test-app/vendor/assets/javascripts", "/Users/example-user/projects/test-app/vendor/assets/stylesheets", "/usr/local/lib/ruby/gems/2.3.0/gems/turbolinks-2.5.3/lib/assets/javascripts", "/usr/local/lib/ruby/gems/2.3.0/gems/jquery-rails-4.1.1/vendor/assets/javascripts", "/usr/local/lib/ruby/gems/2.3.0/gems/coffee-rails-4.1.1/lib/assets/javascripts"]
The nice thing about all of this is easy to miss. It’s the aspect of having conventions. That means that developers—and designers as well, actually—can all have certain expectations about where to look for files and where to place them. That can be a huge (Trump voice) time saver. When someone new joins a project, they have a good idea how to navigate a project right away.
That is not only convenient, but can also prevent questionable ideas or reinventing the wheel all the time. Yes, convention over configuration can sound a bit boring at times, but it is powerful stuff nonetheless. It keeps us focused on what matters: the work itself and good team collaboration.
The assets paths mentioned aren’t set in stone, though. You can add custom paths as well. Open config/application.rb and add custom paths that you want to be recognized by Rails. Let’s take a look at how we would add a custom_folder.
config.application.rb
module TestApp
class Application < Rails::Application
config.assets.paths << Rails.root.join("custom_folder")
end
end
When you now check the search path again, you will find your custom folder being part of the search path for the asset pipeline. By the way, it will now be the last object placed in the array:
Terminal
Rails.application.config.assets.paths
Output
["/Users/example-user/projects/test-app/app/assets/images", "/Users/example-user/projects/test-app/app/assets/javascripts", "/Users/example-user/projects/test-app/app/assets/stylesheets", "/Users/example-user/projects/test-app/vendor/assets/javascripts", "/Users/example-user/projects/test-app/vendor/assets/stylesheets", "/usr/local/lib/ruby/gems/2.3.0/gems/turbolinks-2.5.3/lib/assets/javascripts", "/usr/local/lib/ruby/gems/2.3.0/gems/jquery-rails-4.1.1/vendor/assets/javascripts", "/usr/local/lib/ruby/gems/2.3.0/gems/coffee-rails-4.1.1/lib/assets/javascripts", #<Pathname:/Users/example-user/projects/test-app/custom_folder>]
Concatenation & Compression?
Why do we need that stuff? Long story short, you want your apps to load faster. Period! Anything that helps with that is welcome. Of course you can get away without optimizing for that, but it has too many advantages that one simply can’t ignore. Compressing file sizes and concatenating multiple asset files into a single “master” file that is downloaded by a client like a browser is also not that much work. It is mostly handled by the pipeline anyway.
Reducing the number of requests for rendering pages is very effective at speeding things up. Instead of having maybe 10, 20, 50 or whatever number of requests before the browser is finished getting your assets, you could potentially only have one for each CSS and JS asset. Today this is a nice thing to have, but at some point this was a game changer. These kinds of improvements in web development should not be neglected because we are dealing with milliseconds as a unit.
Lots of requests can add up pretty substantially. And after all, the user experience is what matters most for successful projects. Making users wait for just that little extra bit before they can see content rendered on your page can be a massive deal-breaker. Patience is not something we can expect from our users.
Minification is cool because it gets rid of unnecessary stuff in your files—whitespace and comments, basically. These compressed files are not made with human readability in mind. They are purely for machine consumption. The files will end up looking a bit cryptic and hard to read although it’s still all valid CSS and JS code—just without any excess whitespace to make it readable and understandable for humans.
The JS compression is a bit more involved, though. Web browsers don’t need that formatting. That lets us optimize these assets. The takeaway issue from this section is that a reduced amount of requests and minimized file sizes speed things up and should be in your bag of tricks. Rails gives you this functionality out of the box without any additional setup.
High-Level Languages
You might have heard about them already, but as a beginner you might not be up to speed on why you should learn yet another language—especially if you are still busy learning to write classical HTML and CSS. These languages were created to deal with shortcomings that developers wanted to iron out. They mostly address issues that developers didn’t want to deal with on a daily basis. Classical laziness-driven development, I guess.
“High-level languages” sounds more fancy than it might be—they are rad, though. We are talking about Sass, CoffeeScript, Haml, Slim, Jade, and things like that. These languages let us write in a (mostly) user-friendly syntax that can be more convenient and efficient to work with. All of them need to be precompiled during a build process.
I’m mentioning this because this can happen behind the scenes without you noticing. That means it takes these assets and transforms them into CSS, HTML, and JS before they are deployed and can be rendered by the browser.
Sass
Sass stands for ”Syntactically Awesome Style Sheets” and is much more potent than pure CSS. It lets us write smarter stylesheets that have a few programming tools baked in. What are we talking about here, exactly? You can declare variables, nest rules, write functions, create mixins, easily import partials, and lots of other stuff that programmers wanted to have available while playing with styles.
It has grown into a pretty rich tool by now and is excellent for organizing stylesheets—for big projects I would even call it essential. These days, I’m not sure if I wouldn’t stay away from a large app that isn’t using a tool like Sass—or whatever non-Ruby centric frameworks have to offer in that regard.
For your current concerns, there are two syntax versions of Sass you need to know about. The Sassy CSS version, aka SCSS, is the more widely adopted one. It stays closer to plain CSS but offers all the fancy extensions developers and designers came to love playing with stylesheets. It uses the .scss file extension and looks like this:
SCSS
#main p {
color: #00ff00;
width: 97%;
.redbox {
background-color: #ff0000;
color: #000000;
}
}
Visually, it’s not that different than CSS—that’s why it is the more popular choice, especially among designers I think. One of the really cool and handy aspects of SCSS, and Sass in general, is the nesting aspect. This is an effective strategy to create readable styles but also cuts down a lot on repetitive declarations.
SCSS
#main {
color: blue;
font-size: 0.3em;
a {
font: {
weight: bold;
family: serif;
}
&:hover {
background-color: #eee;
}
}
}
Compare the example above with the respective CSS version. Looks nice to slim down these styles with nesting, no?
CSS
#main {
color: blue;
font-size: 0.3em;
}
#main a {
font-weight: bold;
font-family: serif;
}
#main a:hover {
background-color: #eee;
}
The indented Sass syntax has the .sass file extension. This one lets you avoid dealing with all these silly curly braces and semicolons that lazy devs like to avoid.
How does it do that? Without the brackets, Sass uses indentation, whitespace essentially, to separate its properties. It’s pretty rad and looks really cool too. To see the differences, I’ll show you both versions side by side.
.sass
#main
color: blue
font-size: 0.3em
a
font:
weight: bold
family: serif
&:hover
background-color: #eee
.scss
#main {
color: blue;
font-size: 0.3em;
a {
font: {
weight: bold;
family: serif;
}
&:hover {
background-color: #eee;
}
}
}
Pretty nice and compact, huh? But both versions need processing before it turns into valid CSS that browsers can understand. The Asset Pipeline takes care of that for you. If you hit browsers with something like Sass, they would have no idea what’s going on.
I personally prefer the indented syntax. This whitespace-sensitive Sass syntax is less popular than the SCSS syntax, which might not make it the ideal choice for projects other than your personal ones. It’s probably a good idea to go with the most popular choice among your team, especially with designers involved. Enforcing whitespace hipness with people who have spent years taming all the curly braces and semicolons out there seems like a bad idea.
Both versions have the same programming tricks up their sleeves. They are excellent in making the process of writing styles much more pleasant and effective. It's lucky for us that Rails and the Asset Pipeline make it so easy to work with either of these—pretty much out of the box.
Slim & Haml
Similar arguments can be made about the benefits of using something like Slim and Haml. They are pretty handy for creating more compact markup. It will get compiled into valid HTML, but the files you deal with are much more reduced.
You can spare yourself the trouble of writing closing tags and use cool abbreviations for tag names and such. These shorter bursts of markup are easier to skim and read. Personally, I’ve never been a huge fan of Haml, but Slim is not only fancy and convenient, it’s also very smart. For people who like the indented Sass syntax, I would definitely recommend playing with it. Let’s have a quick look:
Slim
#content
.left.column
h2 Welcome to our site!
p= print_information
.right.column
= render :partial => "sidebar"
Haml
#content
.left.column
%h2 Welcome to our site!
%p= print_information
.right.column
= render :partial => "sidebar"
Both result in the following ERB-enhanced HTML in Rails:
<div id="content">
<div class="left column">
<h2>Welcome to our site!</h2>
<p>
<%= print_information %>
</p>
</div>
<div class="right column">
<%= render :partial => "sidebar" %>
</div>
</div>
On the surface, the differences are not that big, but I never quite got why Haml wants me to write all these extra % to prepend tags. Slim found a solution that got rid of these, and therefore I salute the team! Not a big pain, but an annoying one nevertheless. The other differences are under the hood and are not within the scope of this piece today. Just wanted to whet your appetite.
As you can see, both preprocessors significantly reduce the amount you need to write. Yes, you need to learn another language, and it reads a bit weird at first. At the same time I feel the removal of that much clutter is totally worth it. Also, once you know how to write ERB-flavored HTML, you will be able to pick up any of these preprocessors rather quickly. No rocket science—the same goes for Sass, by the way.
In case you're interested, there are two handy gems for using either of them in Rails. Do yourself a favor and check them out when you get tired of writing all these tiresome opening and closing tags.
gem "slim-rails" gem "haml-rails"
CoffeeScript
CoffeeScript is a programming language like any other, but it has that Ruby flavor for writing your JavaScript. Through preprocessing, it compiles into plain old JavaScript that browsers can deal with. The short argument for working with CoffeeScript was that it helped to overcome some shortcomings that JS was carrying around. The documentation says that it aims at exposing the “good parts of JavaScript in a simple way”. Fair enough! Lets have a quick look at a short example and see what we’re dealing with:
JavaScript
$( document ).ready(function(){
var $on = 'section';
$($on).css({
'background':'none',
'border':'none',
'box-shadow':'none'
});
});
In CoffeeScript, this fella would look like this:
CoffeeScript
$(document).ready ->
$on = 'section'
$($on).css
'background': 'none'
'border': 'none'
'box-shadow': 'none'
return
Reads just a tad nicer without all the curlies and semicolons, no? CoffeeScript tries to take care of some of the annoying bits in JavaScript, lets you type less, makes the code a bit more readable, offers a friendlier syntax, and deals more pleasantly with writing classes. Defining classes was a big plus for a while, especially for people coming from Ruby who were not super fond of dealing with pure JavaScript. CoffeeScript took a similar approach to Ruby and gives you a nice piece of syntactic sugar. Classes look like this, for example:
Class
class Agent
constructor: (@firstName, @lastName) ->
name: ->
"#{@first_name} #{@last_name}"
introduction: (name) ->
"My name is #{@last_name}, #{name}"
This language was quite popular in Ruby land for a while. I'm not sure how adoption rates are these days, but CoffeeScript is the default JS flavor in Rails. With ES6, JavaScript now also supports classes—one big reason to maybe not use CoffeeScript and play with plain JS instead. I think CoffeeScript still reads nicer, but a lot of the less cosmetic reasons to use it have been addressed with ES6 these days. I think it’s still a good reason to give it a shot, but that is not what you came for here. I just wanted to give you another appetizer and provide a little bit of context around why the Asset Pipeline allows you to work in CoffeeScript out of the box.
Final Thoughts
The Asset Pipeline has proven itself far beyond my expectation when it was introduced. At the time it really made quite the splash and was addressing a serious pain for developers. It still does, of course, and has established itself as a frictionless, effortless tool that improves the quality of your applications.
What I like most about it is how little hassle it involves getting it to work. Don’t get me wrong, tools like Gulp and Grunt are impressive as well. The options for customization are plenty. I just can’t shake the comfy feeling that Rails gives me when I don’t have to deal with any setup before getting started. Conventions can be powerful, especially whey they result in something seamless and hassle-free.
No comments:
Post a Comment