How can I run a watir test in the context of the app that's being tested? I'd like my test to browse the app and then access ActionMailer::Base.deliveries for emails or check models directly. This is how I understand what's being described here.
UPDATE: They probably use Capybara to be able to acces the email array and be in the context of the "server" which is instantiated just for the test.
I suggest checking out the Rails unit testing docs, then writing a simple Rails test case that starts your app - then try adding a line or two of Watir code to access your app:
http://guides.rubyonrails.org/testing.html
As far as I know you should be able to write a Rails unit test, then put Watir code inside one of your test methods - and if all goes well you should be able to instantiate your web app, use Watir to launch a browser to test it, and in the same method(s) perform non-Watir low-level testing (e.g. checking models/data/etc.)
I've never used Watir inside a Rails test, but I don't see why it wouldn't work.
Watir is about driving browsers to automate functional testing. You could I suppose use it for unit testing of the top level UI stuff, but more often in a 'unit test' context that would be done using a headless browser emulation, Capybara, celerity, or watir-webdriver using the headless option, purely for speed of operation since driving an actual browser can be slow even with a fast browser like chrome.
Most of the times people use Watir it's for more functional tests, often from a test runner framework like Cucumber, sometimes Rspec depending on your needs. You might combine that with other ruby code to access or create test data, to validate something made it into the DB from the UI, but everything in the Watir gem is all about the browser and interacting with it much like a human would, and driving the browser is it's function within the set of tools you might use.
I had the same need and found the following solution: https://stackoverflow.com/a/9687562/90741. I reproduce it here, as the linked question seems to be dead, with its owner...
I had the same need (Watir+RSpec), but the problem is: there is no Rack stack running by default during tests...
I found this solution:
In spec_helper.rb:
HTTP_PORT = 4000
test_instance_pid = fork do
exec 'unicorn_rails -E test -p %d' % HTTP_PORT
end
at_exit do
Process.kill "INT", test_instance_pid
Process.wait
end
Which start the test stack once for all spec tests, and kill it at the end. (In this sample, I am using unicorn, but we can imagine using any other server)
In the spec, I reuse the HTTP_PORT constant to build URL:
browser.goto "http://localhost:#{HTTP_PORT}/"
Related
I have a Rails app that interacts with some Mac software and I need to write some tests for it. How on earth do I do that? Where do I even start?
The Rails app connects to the Mac app through AppleScript and Terminal. Any ideas?
Update
Found this gem to help with Shell expectations. Is that as good as I'm gonna get?
https://github.com/matthijsgroen/rspec-shell-expectations
Testing external dependencies can certainly be a challenge.
Remember that tests you write should test the behavior of the Rails app, not the external dependency. You should have integration tests to verify that the code actually works with the application, but they should be "smoke tests", not a full suite to test every feature.
Write unit tests to verify the behavior of the code that relies on the dependency, and mock out the interactions. Typically with command line apps that means:
the app wrote to STDOUT
the app wrote to STDERR
the app read from STDIN
the app exited with a particular status code
The gem you mentioned is a good start, but you may find it worthwhile to look at rolling your own helper code using Open3 from the Ruby standard library, which can be useful for all the items in the list above.
Use a tag on the specs that need to use the Mac application to make it easier to filter out those specs as necessary.
You may already be familiar with vcr for mocking out HTTP interactions; its "playback" feature is a good source of inspiration.
We have a rails app and some users have used Selenium IDE firefox extension to record/create some tests. We'd like to be able to integrate these tests into our codebase of automated tests.
I understand that Selenium IDE has the ability to export test cases, but what is the easiest way to convert these tests into something that rspec can run without a browser (or headless)?
I've used both. Here is what you can do
export the file using the rspec webdriver format
adjust any use of variables or local file storage that is possible in selenium to make sense using rspec. This will depend totally on your usage.
Here's what one goes through:
Select the test you want to export (you can also export the entire suite)
Export as rspec webdriver
Open the resulting file
Adjust as needed, e.g. move common items to spec_helper.rb and make sure it is included.
Put the file in an rspec folder, e.g. spec/views and make sure that spec_helper includes them.
I understand that Selenium IDE has the ability to export test cases, but what is the easiest way to convert these tests into something that rspec can run without a browser (or headless)?
Selenium IDE is only plugin for Firefox.
I believe, you can to do the same steps on another language/way.
But for this you need to have another code.
Code for Selenium Web Driver is possible to use with headless browser as phantomJS. And its easy.
But Selenium IDE code is only for Firefox IDE plugin.
And by the way, its hard to maintain tests written using IDE. I believe, better to rewrite all tests using Selenium WebDriver in your case. Its will be much easier and faster. (you can use Page Object model for this, as example)
I'm working on an advertising application, where there is some cross-domain JS. I'd like to write tests that verify that the cross-domain JS is working as intended.
My first thoughts are that I would need to be able to
visit some_url_thats_not_my_rails_app
However, Capybara throws a "No Route Matches", since it obviously goes to only relative paths, and is intended for testing your OWN website. But I really need to go on ANOTHER page, and verify that things like the serialized token are identical.
Is Capybara the right tool for this? If so, what do I need to do to force non-relative paths?
Yes, Capybara is right tool for it. You can also use Ruby with Selenium webdriver(Would be awesome with page object gem) or Ruby with Watir webdriver.
To visit and test any web application, you can use Ruby and Capybara. For this you need to set app_host. Add this one in support/env.rb:
Capybara.app_host = "http://flipkart.com"
And in hooks.rb
Before do
visit('/')
end
You can use Capybara with Selenium webdriver and you will be good to test any deployed application in production, staging or prep env.
I started a blank project with methadone, an awesome framework for building command line apps. The only problem is that I am unable to debug from within the App class that's in bin/my_app
The App class is a file created when you run methadone. Here's how I'm trying to use pry
#!/usr/bin/env ruby
require 'optparse'
require 'methadone'
require 'my_app'
require 'pry'
class App
include Methadone::Main
include Methadone::CLILogging
main do
binding.pry # <= pry here
end
...
When I run rake features I can tell the running process is trying to do something with pry since it pauses for a few seconds. I get the following error and then rake/cucumber is aborted.
process still alive after 3 seconds (ChildProcess::TimeoutError)
I can use pry just fine from the cucumber steps, rspec, or any other place, just not from anywhere in this App class.
One very interesting thing is that if I run my command line app from the console it WILL stop where pry is. It just wont pop into pry when using cucumber.
How can I get the app to work with pry when I'm running rake features?
Update
Sorry, I should clarify that methadone comes with aruba. So my cucumber scenario would look like this
When I successfully run `my_app arg1`
However, it WILL go into debug/pry if I run it with
bundle exec bin/my_app
Use pry-remote to connect to a pry session in the Aruba-managed subprocess.
(Disclosure: I paired with #Dty to come to this solution)
Aruba runs the app in a totally separate process, so I would guess what's happening is that when aruba runs your app, pry starts up at a prompt and waits for input. Since it doesn't get any, aruba times out after three seconds (the default it will wait for an app to complete). This is why you see the "process still alive" issue.
I'm not 100% sure how you could get the standard input of your shell that's running rake features to connect to your app's standard input so you could issue pry commands, but I don't think aruba was designed to allow this.
You have a couple of options:
Tag your scenario with #announce, and use When I run interactively... followed by several When I type - these commands should go to the interactive pry console that's waiting. Kind of kludgy, but it might work
Execute a unit test of your App class. You'll need to replace the call to go! with something like go! if $0 == __FILE__ so that you can require your executable in a test and manipulate App directly.
I have not tried either of these, but the second option feels a bit better and could also be improved with support from the library, if you can figure out a good way to do this.
In this video from GoGaRuCo 2011, Corey Haines shows some techniques for making Rails test suites much faster. I would summarize it as follows:
Put as much of your code as possible outside the Rails app, into other modules and classes
Test those separately, without the overhead of loading up Rails
Use them from within your Rails app
There were a couple of things I didn't understand, though.
He alternates between running tests with rspec and spn or spna (for example, at about 3:50). Is spn a commonly-known tool?
In his tests for non-Rails classes and modules, he includes the module or class being tested, but I don't see him including anything like spec_helper. How does he have Rspec available?
Sorry about the confusion. spn and spna are aliases I have that add my non-rails code to rspec's load path. There isn't anything special about them, other than adding a -I path_to_code on the command-line.
These days, I add something like this to my .rspec file:
-I app/mercury_app
Then I can do simple require 'object_name' at the top of my specs.
As for not including spec_helper: that is true, I don't. When you execute your spec file with rspec <path_to_spec_file>, it gets interpreted, so you don't need to require rspec explicitly.
For my db specs these days, I also have built an active_record_spec_helper which requires active_record, establishes a connection to the test database, and sets up database_cleaner; this allows me to simply require my model at the top of my spec file. This way, I can test the AR code against the db without having to load up my whole app.
A client I am working at where we are using these techniques is interested in supporting some blog posts about this, so hopefully they will start coming out towards the middle of June.