Regression Test Automation Framework

Quick Start

• Install Git.
• Pull down the repository to your local computer (git clone).
• Get Ruby setup.
  • Windows: Install Ruby 3.1.3.
  • OSX: Install Ruby Version Manager (RVM).
  • OSX: Install the appropriate version of Ruby using the version specific install command rvm install ruby-X.X.X. (Recommend 3.1.3.)
• gem install bundler if bundler is not already installed.
• bundle to install required gems. (You need to run this whenever pulling down this repo with new changes.)
• rspec spec/signup_spec.rb to run a single test and verify it's working.
• rspec to run all tests.

Frameworks

Page-Object

We're using the page-object gem to wrap around Watir. You can always drop down and use the Watir browser object, but page-object provides some nice abstractions.

• Elements
• Simple DSL
• Ajax Calls

Watir

An open source Ruby library for automating tests. Watir interacts with a browser the same way people do: clicking links, filling out forms and validating text. The docs and examples can be found at https://watir.github.io/. Most things you need to do will be handled using the page object methods (see above), but if you need the browser to do something specific that page objects can't do you can call it directly with @browser.<command>.

faker

The Faker Ruby gem can be used to generate random fake data within tests.

RSpec

• RSpec Expectations syntax

Writing Tests

Structure

• lib/pages contains objects that represent a web page. Example: home_page.rb maps to the home page for EZ Contact Manager.
• spec/ contains the tests.

Rules

1. A given lib/pages/whatever_page.rb file should map to a single web page.
2. Test in Watir the absolute minimum that's required since these tests are slow and more brittle than hitting the database directly.

Things to consider when writing tests

1. We've run into several scenarios where workflows are very similar or an abstraction could be made. We decided simplicity, readability, and accessibility are more valuable than conciseness in many cases. If you have a series of tests with a bunch of if conditions trying to determine what state the tests need to run, it's probably better to duplicate that test for each mode.