Appium: Ruby: Distributed tests with appium for iOS - ios

Question

I am having a fully fledged suite for automated tests written using ruby with Appium for mobile automation.
I am running these suites in one simulator in one machine and it takes up a lot of time, around 1 hour to run 56 test cases(We have system test cases where multiple checks like database/Api/functional are integrated). We have more additional test cases adding to our way.
We have implemented running our tests across 3 mac machines currently with running different cucumber tags integrated to Jenkins. However, more addition of tests is only going to take us more time or more mac's
With xcode 9 we can initiate multiple simulators on one machine at the same time, and I had like to know, if there is any sample scenario or documentation on how to implement distributed tests across simulators in one mac machine
I had tried loading two or three different desired capabilities with different platform version, but it only loads the tests in sequential order.
I had gone through a lot of material online that has only the steps to make this possible in android. Does iOS support it?
Or could anyone possibly provide links that would help me? to
1. Implement distributed tests across various simulators in one mac
2. Use cucumber tags to distribute tests creating instance for each desired capability
Update:
I had tried implementing the multithread option and tried to initiate the tests to specific simulator creating instance with each thread. However, I find the tests not running in parallel but sequential.
This is my code:
def start_app(device_id, wdalocalport)
caps_config = {
platformName: "iOS",
wdaLocalPort: wdalocalport,
deviceName: "iPhone Simulator", #update device as per your need
app: (File.join(File.dirname(__FILE__), "Sequoia.app")),
bundleId: "something",
automationName: "XCUITest",
xcodeOrgId: "something",
xcodeSigningId: "iPhone Developer",
#platformVersion: "10.2",
noReset: "true",
fullReset: "false",
showIOSLog: "true",
autoAcceptAlerts: "true",
showXcodeLog: "true",
useNewWDA: "true",
resetOnSessionStartOnly: "true",
udid: device_id }
appium_lib_config={ port: 4723 }
$opts={ caps: caps_config, appium_lib: appium_lib_config }
setup
end
def setup
#appium = Appium::Driver.new($opts)
#appium.start_driver
#Makes all appium_lib methods accessible from steps
#Starts appium driver before the tests begin
end
def test(device1,device2)
threads = []
threads << Thread.new {
start_app(device1, '8100')
}
threads << Thread.new {
start_app(device2, '8200')
}
threads.each(&:join)
end
end
I am calling the launch tests using the test method passing the udid's. The simulators launch at the same time, and also installs the application at the same time, but the tests aren't parallel.
Any help to improvise this case?
I was able to use the rake to parallel run, but I still find this approach runs the tests in sequential manner or doesnt run at all
PFB the code
def run_cucumber(cucumber_options)
Cucumber::Rake::Task.new do |t|
t.cucumber_opts = cucumber_options
end
Rake::Task[:cucumber].invoke
end
task :iphone_7 do |t|
ENV['DEVICE'] = 'iphone7'
run_cucumber('-r features features/test.feature --format pretty --tags #slave1')
end
task :iphone_8 do |t|
ENV['DEVICE'] = 'iphone8'
run_cucumber('-r features features/test.feature --format pretty --tags #slave2')
end
multitask :all => [:iphone_7,:iphone_8]
My hooks.rb
Before do
check
end
def check
if ENV['DEVICE'] == 'iphone7'
start_app('iPhone6','port','udid')
elsif ENV['DEVICE'] == 'iphone8'
start_app('iphone6','port','udid')
else
puts "Device not"
end
end
I have been getting DEVICE NOT. Not sure what am I missing.

Answer 1

You need to create a method which invokes cucumber rake first in the Rakefile as follows:
def run_cucumber(cucumber_options)
Cucumber::Rake::Task.new do |t|
t.cucumber_opts = cucumber_options
end
Rake::Task[:cucumber].invoke
end
This method will need cucumber_options. You can pass same options as you pass on the command line like this:
-r features features/test.feature --format pretty --tags #test1
After this, define rake tasks for a particular configuration in the same Rakefile:
task :iphone_7 do |t|
ENV["DEVICE"] = "iphone 7"
run_cucumber('-r features features/test.feature --format pretty --tags #test1')
end
task :iphone_8 do |t|
ENV["DEVICE"] = "iphone 8"
run_cucumber('-r features features/test.feature --format pretty --tags #test2')
end
Here, ENV["DEVICE"] is used set to device config. You can use this variable in your env.rb in Before block to initialize the test. In Before block you can initialize devices based on this environment variable.
You can create rake tasks like this as many as you need.
Then you can invoke multiple tasks at the same time by using ruby's parallel gem. Examples are in this link -> Ruby Parallel gem
Just pass different types of rake task as the arguments in the Parallel object.
Eg:
task :run_parallel do |t|
Parallel.map([:iphone_7,:iphone_8]) {|task| Rake::Task[task].invoke }
end
And then trigger the run_parallel rake task to execute the whole code.
Eg:
bundle exec rake run_parallel