You want the big deal? You can’t handle the big de—err. Sorry. I think I just got a little bit, ahem, carried away. Not using the console results in RSI from constant page reloading, self-inflicted hair loss, and a distinct rise in cortisol.
But on the positive side of things, the console lets you do things like this without faking out your controller and hitting reload a bunch of times:
>> @george = Person.find_by_name('George') >> @bob = Person.find_by_name('Bob') >> @bob.friends << @george
Now, I’ve left out the return values of all these actions, for readability’s sake. But every time you perform a line of code in the console, you’ll get back the result—it’ll show you how the object has changed, or whether an action resulted in true or false, and so on. You can inspect and manipulate stuff, live, without writing the code in a controller, saving the file, reloading your browser, etc., etc., ad nauseum.
It’s not only a time saver, it’s a much better development practice.
When you create a new Rails app (project directory), it comes complete with a script
directory, which includes, among other things, the generate
script which can create controllers, models, and other files for you (including scaffolding).
But we’re not talking about generate
today, we’re talking about the console . As in, script/console
. Which is not only its name, but how you run it. Take up your terminal or command-line, pop into your Rails app’s root directory (aka RAILS_ROOT
) and take a whack at the following:
$ script/console
Ding! You will be rewarded with something like the following:
Loading development environment. >>
The first line will tell you which Rails environment your console is running in—and yes, you can specify it (see below). The second line is your prompt—followed by a fat little text insert bar which, depending on your OS and shell configuration, might be blinking, or might not. But either way, it’s beckoning to you, “Play with me!” (the little flirt).
By default, script/console
will load itself using your development production environment—unless your environment variables specify otherwise.
You can also tell the console at run-time which environment to start in, simply by tacking the environment name onto your command:
$ script/console production $ script/console test $ script/console development
When you’re in the console, you have access to all your Rails app’s models and other classes, in addition to all the goodies that the Ruby standard lib offers.
Whenever you enter an expression into the console, the result of that expression will always be displayed, preceded by the =>
marker:
>> "monkey" + "ball" => "monkeyball" >> 6 * 7 => 42
In the case of complex objects, this can be a messy thing indeed:
>> @bob = User.new({:name => "Bob", :job => "Test Dummy"}) => #<User:0x2645874 @new_record=true, @attributes={"name"=>"Bob", "job"=>"Test Dummy"}>
You can enter code into the console just like you would in a file, or another terminal-based application:
>> ['foo','bar','bat'].each { |word| puts word } foo bar bat => ["foo", "bar", "bat"]
Or you can enter code involving multiple lines:
>> ['foo','bar','bat'].each do |word| ?> puts word >> end foo bar bat => ["foo", "bar", "bat"]
The ?>
indicates that you’re inside a block of code which will need to be terminated one way or another.
If you’re in an indented block of code (e.g., something that should be terminated with end
) and want to escape without finishing, you can use ctrl + c
to get back to the top level. The current expression will be ignored and not run.
Oops! You wrote some fancy code in the console and hit Enter
—but you forgot to actually store it in a variable. But all is not lost, and you don’t have to run the code again. The console provides a magic variable ( _
) for your use, which will automatically give you the last return value the console saw:
>> "monkey" + "ball" => "monkeyball" >> string = _ => "monkeyball" >> string => "monkeyball"
You can perform operations directly on _
, but remember that that will modify the contents of _
as soon as it’s run.
Just like in most terminal applications, you can access your command history using your keyboard’s up and down arrows. Up goes backwards in time… down, of course, goes forward (and you can only go forward when you’re already backward—so much for irb
being the secret to time travel. sigh!).
There are a couple ways to check out objects in the console. The first, of course, is to just make use of the way it shows the inspect
method on everything:
>> @bob => #<User:0x2645874 @new_record=true, @attributes={"name"=>"Bob", "job"=>"Test Dummy"}>
But there’s a much more readable alternative:
>> puts @bob.to_yaml --- !ruby/object:User attributes: name: Bob job: Test Dummy new_record: true => nil
But wait! There’s more! Or, less, really, because there’s a much shorter way:
>> y @bob
YAMLicious.
You can require libraries in the console just like you can in a file containing Ruby code:
require 'mylib'
You may have to provide a file path if the file is not in one of the places that the console looks.
When you change your models and such and need to see those changes reflected in your Rails app in a web browser you just hit Reload
. In dev mode, the Rails app loads all the files anew every single time a page is requested. The same is true for the console—but it loads the files anew every time you start the console, instead of per page request.
To avoid having to quit and reopen the console every time you change your application’s code, simply type the follow command:
>> Dispatcher.reset_application!
Or, better yet:
>> reload!
This is what you’ll see:
>> reload! Reloading... => [ApplicationController, User]
All your models and suchlike will show up in this array, showing you what is being loaded. The best part is, you won’t have to refetch your objects—they’ll be updated.
Just type exit
. Yep, it’s that easy.
The console is ideal for working on your models, for such tasks as:
It’s also a great way to become more comfortable writing Rails code off the cuff… it’s easier to practice if the process isn’t as tedious.
A note for the cowboy coder in all of us: The console is not a replacement for writing unit tests. It’s a complementary technique. Write those tests, you naughty thing.
You can use it for non-model stuff, as well, but it’s a little bit trickier:
irb
—and console, specifically.If you have tips for how to use the console, please gimme comments!
Here’s a handful of things you might not know about the console, or things you can do to make your console experience better. (The “making it better” stuff typically is done through supplementing Ruby’s irb
—interactive Ruby—console, which the Rails console is based on. You can place arbirtrary Ruby code into irb
’s initialization script to affect the way it’s used.)
The console has baked-in autocompletion functionality.
Start typing a class name, and hit tab—it will either autocomplete it, or offer you a list of choices:
>> Str String StringInput StringScanner StringIO StringOutput Struct
It works for Ruby standard lib classes and Rails built-in classes (e.g., Dispatcher, ActionController, and so on), but not for the controllers and models in your app/
directory.
You can also use autocomplete for method names, on any class or object, regardless of where it’s from—so long as it’s within the console’s purview:
>> User.a User.abstract_class User.after_update User.abstract_class= User.after_validation User.abstract_class? User.after_validation_on_create User.accessible_attributes User.after_validation_on_update
(And so on. I cut the list short because it’s ridiculously long.)
If hitting tab after a class name and a period (e.g. User.[tab]
) doesn’t work the first time, hit it again—the console is trying to save you from making a grievous mistake which results in a list of hundreds of methods. The second time you hit tab, it will ask you:
>> User. Display all 354 possibilities? (y or n)
Hitting n
will return you to the line you were typing, and y
will, of course, display them all. Use your space bar to page through the long lists, or Return
to go line by line.
I don’t know about you, but I’m neurotic about using my terminal windows—I hate it when I’m constantly working at the bottom of the window, or there’s a huge swash of unnecessary output above where I’m working.
Unfortunately, the typical bash clear
command doesn’t work in the console:
>> clear NameError: undefined local variable or method 'clear' for # from (irb):1
But ctrl + l
(that’s a lowercase L) will do the trick. For Mac users, Command + k
might be more comfy to type. (Thanks Phil !)
You can turn on vi-style commands by editing your user account’s .inputrc
file:
$ cat ~/.inputrc set editing-mode vi
This will also affect any other console-type tools which use the readline
library. Which is probably a happy side effect, but you’ve been warned!
err the blog has an article called irb Mix tape which describes, among other really useful things, how to run the console in a sandbox—meaning that all your database changes are completely reverted when you exit the console.
Basically, it looks like this:
$ script/console --sandbox Loading development environment in sandbox. Any modifications you make will be rolled back on exit. >>
You can also use -s
as a flag instead of the more verbose --sandbox
. Sweet!
When you exit the console, your history (accessed by those keyboard arrows) is wiped clean. But it doesn’t have to be that way. The irb Tips and Tricks page at RubyGarden shows you how you can make that history stick around.
See the Further Resources list for a bunch of links showing how to modify the console / irb
to make ‘em better.
Like this article? Digg it … and consider dropping a few coins in the tip jar .
Here are all the links I mentioned earlier in the article, and then some:
ri
commands, etc., without editing the ~/.irbrc
file) People:
irb
wizardry now and then