kronosapiens.github.io - Getting Cozy with rspec_helper

Getting Cozy with rspec_helper

Kronosapiens.github.io Spined HTML

Getting Cozy with rspec_helper AbacusWell-nighGetting Cozy with rspec_helper Mar 12, 2014 ZOMG THIS IS A LOT OF CODE ALL OF A SUDDEN WTF Relax. This is rspec_helper, or ‘Specky’ for short. You’ve met. In fact, you two have once spent quite a bit of time together. Quality time. It’s not unchangingly an easy relationship, of course. You have your differences, but you’re on the same team. You want Specky to help you run your test suites, to requite you the tools to run constructive and efficient tests. Specky wants you to unmistakably communicate your desires and expectations. Problems upspring when you don’t know what Specky is trying to tell you, and when Specky doesn’t know what you want. As with all relationships, it comes lanugo to communication. ENV["SINATRA_ENV"] = "test" require_relative '../config/environment' require 'rack/test' require 'capybara/rspec' require 'capybara/dsl' if defined?(ActiveRecord::Migrator) && ActiveRecord::Migrator.needs_migration? raise 'Migrations are pending run `rake db:migrate SINATRA_ENV=test` to resolve the issue.' end RSpec.configure do |config| config.treat_symbols_as_metadata_keys_with_true_values = true config.run_all_when_everything_filtered = true config.filter_run :focus config.include Rack::Test::Methods config.include Capybara::DSL config.before(:suite) do DatabaseCleaner.strategy = :transaction DatabaseCleaner.clean_with(:truncation) end config.around(:each) do |example| DatabaseCleaner.cleaning do example.run end end config.order = 'default' end def app Rack::Builder.parse_file('config.ru').first end Capybara.app = app In interests of peace and harmony, I thought it was upper time to sit the two of you lanugo for a little coeur-a-coeur. What, exactly, is Specky all about? Let’s find out. ENV["SINATRA_ENV"] = "test" What’s going on here? As you probably noticed, most of your projects have two databases going on – one for development, and one for testing. The way that your program knows what database to write and read from is by checking the “SINATRA_ENV” variable.ConsideringSpecky is run surpassing every rspec file, this line will make sure that all your tests are writing and reading from your test database – just the way you want them to. require_relative '../config/environment' require 'rack/test' require 'capybara/rspec' require 'capybara/dsl' Here, Specky is doing some housekeeping and loading all the other files you’ll need to run your tests. The first line is the big one – this loads all your models, your controllers, your views – in short, everything in your program’s environment. The next lines load the files needed for testing – the rack/test and the capybara methods, so you can enjoy the oh-so-convenient testing syntax you’ve come to enjoy (all those “describes” and “expects” need to come from somewhere, without all). if defined?(ActiveRecord::Migrator) && ActiveRecord::Migrator.needs_migration? raise 'Migrations are pending run `rake db:migrate SINATRA_ENV=test` to resolve the issue.' end Here, Specky is double-checking your work and making sure that you’ve set up your testing database the way you’re supposed to. Specky gets around, and they know that developers have a bad habit of creating migrations, running downstairs for some nutrition coke, and forgetting to unquestionably run the migrations to set up their databases. It stresses Specky out when tests explode considering the databases aren’t right, and so they do a quick trammels surpassing they start running your tests. A stitch in time saves nine, as they say. RSpec.configure do |config| config.treat_symbols_as_metadata_keys_with_true_values = true config.run_all_when_everything_filtered = true config.filter_run :focus Specky just opened the big configuration method, and is well-nigh to start flipping some switches. These next three settings are a bit obscure, and unless you’re a pretty sophisticated rspec user,  most likely won’t stupefy the outcomes of your tests. They relate to a full-length of rspec that we don’t really use, in which you set “tags” on specific tests, so that you can run only unrepealable blocs of tests in your suite. If you’re interested in learning increasingly well-nigh these functionalities, the relevant documentation can be found here and here. config.include Rack::Test::Methods config.include Capybara::DSL These next two lines aren’t particularly groundbreaking. They’re the corollaries of the older rack and capybara requirements; this time around, we’re including particular modules into our rspec configuration (as opposed to incorporating full libraries into our environment, which we did before, and which was a necessary precursor to these lines of code). config.before(:suite) do DatabaseCleaner.strategy = :transaction DatabaseCleaner.clean_with(:truncation) end Ah. Now things start to get dicey. Here we move past global rspec configuration settings and start talking well-nigh specific procedure that Specky will execute during the process of running your test suite. This particular quartet requires a little linguistic precision. We often use the term “suite” to describe our unshortened set of tests, as in “running your test suite”. Specky, however, is possessed with a shade increasingly nuance. They use the term “suite” to refer to a woodcut of tests contained within a describe woodcut (at least, equal to this corner of StackOverflow). Describe blocks occupy the second tier in rspec land, stuff used often as the envelope for a number of “it-do” blocks. What this lawmaking is doing substantially is transplanting out your databases in between “suites” (in this case, “transaction” seems to indicate a “rolling back” of data-writing, while “truncation” seems to be increasingly of a “delete everything” sort of deal – see StackOverflow for increasingly on this). For your purposes, though, this ways that any data you create within one test suite will closure to exist by the time you reach your next test suite. This requires a bit of sustentation to precisely where and how you create the data you wish to test; inattention here will result in quite a bit of frustration as you try to debug your objects – only to realize that you are in fact debugging your test suite. Moving on. config.around(:each) do |example| DatabaseCleaner.cleaning do example.run end end This lawmaking was tricky. Specky was stuff quite coy (koi?), here. To be honest, I’m really not 100% on this – the internet was as zipped as Specky was. It’s not entirely well-spoken to me what “example” is, although previous lines would suggest that it is a “suite”, pursuit our precise definition. Then it seems like spare cleaning is done, although given that our previous permafrost of lawmaking seemed to basically obliterate any extant database-type data-oids, I’m not quite sure what this is meant to be doing. Let’s say it’s doing some deep cleaning? Levity aside, the warlike data-clearing on the part of rspec_helper is very important for our tests. In production, you’ll have little tenancy over the data coming into your database. If you were to run your tests with a database containing old data, data that you had created yourself, data from past tests, etc, you run the risk of passing tests not considering of the quality of your program, but by virtue of some fluke of your database. Specky’s idea here is that your databases should contain the minimum value of information necessary to make your tests pass, and so works nonflexible to make sure that every test suite is run versus a figurative tabula rasa. config.order = 'default' end Ah yes, the end of our epic “config” call. And finally – some stability! This line sets the order of tests to something nice and sane, like ‘default’, in which everything runs in the awfully-comfortable order in which it was defined. Compare this to the frustrating-but-admittedly-rather-useful “config.order = ‘random’”, which randomizes tests to add flipside layer of rigor to the overall suite. def app Rack::Builder.parse_file('config.ru').first end In this little trio, Specky will parse your config.ru file using the Rack::Builder.parse_file method, nab the first entry* in the ensuing hash (which is, in fact, the thing that you want) and set that puppy equal to “app”, for use very shortly. Capybara.app = app Finally! The last line. Here, Specky takes the variable “app” which you specified not three picoseconds prior and sets it as the value of the Capybara.app variable – thereby teaching Capybara how to mimic your application. And, voila. The thoughts, dreams, and aspirations of your rspec_helper file have been laid yellowish surpassing you. Hopefully now, in your new and increasingly intimate understanding, the two of you can move bravely forward into a unvigilant new dawn of pleasant, test-driven development. *For the curious, here is what ‘app’ is set equal to: https://gist.github.com/kronosapiens/9505813 Comments Please enable JavaScript to view the comments powered by Disqus. Abacus Abacus kronovet@gmail.com kronosapiens kronosapiens I'm Daniel Kronovet, a data scientist living in Tel Aviv.