I want to change the cucumber reporting to serenity reporting. I am using cucumber 2.3.1 version in our framework. What are the steps i need to start with. Is it possible? As of now, i started adding the dependencies for serenity gradle plugin and replaced cucumber "io.cucumber:cucumber-spring, cucumber-java, cucumber-junit" with "serenity-core, serenity-junit, serenity-screenplay" but some build issues are coming. Not sure if this is doable or how should i approach for this. Is there any example project available?
It is very much doable. To use Serenity, we need either Cucumber or JBehave. So, it is good that you are already using Cucumber. You need to add Serenity dependencies and Cucumber one like serenity-core, serenity-junit, serenity-cucumber5, cucumber-java, cucumber-junit and serenity-screenplay in POM.xml.
The test Runner for Serenity Framework will be something like this.
import org.junit.runner.RunWith;
import io.cucumber.junit.CucumberOptions;
import net.serenitybdd.cucumber.CucumberWithSerenity;
#RunWith(CucumberWithSerenity.class)
#CucumberOptions(plugin = { "pretty" }, features = "src/test/resources/features/LoginPage.feature")
public class CucumberTestSuite {
}
There is a article which explains how a project in Serenity can be built. But it is in Maven. Still you will get some idea. Here, it is link.
Related
Today I've got some theoretical question. I have a little experience in BDD with Cucumber. Now I started working in API testing and I faced with Serenity BDD framework on my new project. It is used with Cucumber. So I would like to clarify what is the difference between Serenity BDD Framework and Cucumber and why can't we use just Cucumber for our tests, because it seems to me, like these two tools are doing the same job. Could anybody give me some explanation or maybe some link to correspondent documentation. Thanks in advance!
No, those are 2 different things.
Cucumber is the layer to map BDD syntax, which is written in .feature file, with actual code that does the job.
Serenity BDD is the framework supporting 3 different approaches:
Cucumber: same features as stand-alone cucumber, can work with UI or API Automation
Page Object: works with UI automation (selenium)
Screenplay: a design pattern for UI and API automation
Serenity BDD does many things for auto testers that:
Config-oriented: serenity.properties or serenity.conf. For example: take screenshot when FOR_EACH_ACTION, BEFORE_AND_AFTER_EACH_STEP, AFTER_EACH_STEP, FOR_FAILURES, DISABLED. This is really helpful for debugging. https://serenity-bdd.github.io/theserenitybook/latest/serenity-system-properties.html#_serenity_take_screenshots
Living report: much more better comparing to cucumber. https://serenity-bdd.github.io/theserenitybook/latest/living-documentation.html
Wrap other libs in nice and clean APIs: serenity-appium, serenity-browserstack, serenity-cucumber, serenity-rest-assured, serenity-saucelabs, serenity-shutterbug1x ... You see all of them at https://github.com/serenity-bdd/serenity-core
I can see in many cases that people prefer Junit rather than TestNg when they are integrated with the cucumber framework but I never understand the reason and benefit of using JUnit over TestNG.
If anyone knows the exact reason please help me to understand. Also, most of the tutorial videos are using Junit.
Can anyone please help me in understanding what is difference or advantages of using Cucumber+TestNG over Cucumber+Junit ?
TestNG provides more Annotations Like #BeforeSuite .
TestNG it self provides support for Junit.
With Latest upgrade of cucumber(4.0) Cucumber - TestNg provides Parallel Execution at Scenario Level where as Junit provides at Feature level.
TestNG and Junit are just a unit testing frameworks, when you use Cucumber with Junit you can use Junit Annotation/Assertions in step definition file and similarly when you use Cucumber with TestNG you can use TestNG Annotation/Assertions in step definition file.
Actually in both cases you would be writing the Test Steps using selenium and probably will use the classes inside JUnit or TestNG to run the cucumber Feature file.
As per my knowledge the only place I used such file was the when I'm creating the TestRunner.java class.
It's much easier to do with JUnit since you just have to import org.junit.runner.RunWith But if you are using TestNG you'll need to use dataProviders and go on a bit more complex path to create the TestRunner Class
Other than that there won't be any other differences since you will be using cucumber features for testing.
Due to the above mentioned advantage, I recommend to use JUnit + Cucumber + Selenium for your scripting
I tried both but decided for JUNIT 4.12 (JUNIT 5 does not seem to work with cucumber yet). It has all that you need. But i have to test that parallel thing in Test NG i guess. ;)
I use Cucumber in a Spring Boot Project and therefore JUNIT did not cause any problems for me. I can hand over a browser instance over multiple classes here (#Autowired) and this makes my code so much more modularizable.
You may have more annotation-options with Test NG but somehow i have the feeling that more people work with JUnit, so you have more support and a bigger community here.
I am trying to integrate BDD using Cucumber. But I am really confused what is the difference between io.cucumber and info.cukes libraries. And which one to use and when.
I tried to read and understand the github README.md file still can't make heads or tails.
Still further I am not sure what is cucumber-jvm. Why do we need cucumber-junit (can't the standalone junit library suffice).
Thanks in advance. Any help is much appreciated.
Refer to the release notes for more details. - https://github.com/cucumber/cucumber-jvm/blob/master/CHANGELOG.md.
There has been substantial changes in cucumber 2. Refer to this for more - https://cucumber.io/blog/2017/08/29/announcing-cucumber-jvm-2-0-0
io.cucumber and info.cukes are Maven group ids. info.cukes was for Cucumber version till 1.2.5. The latest version are in io.cucumber starting from 2.0.0. There is also a new version 3 with more goodies in github with the master as mentioned in the release notes.
The reason the groupid was changed because gherkin has changed the groupid similarly.
cucumber-jvm is the java implementation of Cucumber framework. there are many other implementations in other languages - https://github.com/cucumber.
When you use the #RunWith(Cucumber.class) on top of the test class, it means that a specialized runner is being used which will execute the feature files. The default runner of junit will not get you anywhere, though might cough up some exceptions.
I have inherited a Java / Maven / Cucumber project. I am fairly new to Cucumber.
Inside one of the folder I have a class like this...
import com.intuit.karate.junit4.Karate;
import org.junit.runner.RunWith;
#RunWith(Karate.class)
public class RoadsRunner {
}
Then in the same subdirectory / package I have a .feature file.
with a number of scenario's.
Feature: Check transaction
Background:
* url apiHost + '/api/v1'
* configure headers = {'X-TransactionID': '#(Math.random().toString())' }
Scenario: Get Classes
# get classes
Given path '/myUrl/classes'
And param processName = 'myProcess'
When method get
Then status 200
Question One.
I am using Eclipse. Is there a way I can debug through the test in a similar way that I would debug a Java app?
I have downloaded myself the Cucumber Eclipse plugin but can't quite figure out how to use it.
Question Two.
Without using a custom plugin to debug is there anything I can add to the scenarios to maybe print extra debug information.
thanks
The Cucumber Eclipse plugin gives you 2 things:
IDE syntax coloring / formatting support
Being able to right-click and run a Feature directly without the JUnit "runner"
Karate is Java behind the scenes so you can debug and set break-points, but it may not be as seamless as you expect. In 0.6.0 you have the option of placing a conditional break-point in Karate code that runs before / after each test step - see screen-shot.
So as you rightly called out, printing to the log might be the most effective way to work through complicated test scripts. Please refer to the print keyword - which is exactly what you are looking for.
2 more points:
the optional HTML report includes all HTTP request / response logs - which is great for troubleshooting a test.
I would love for the Karate UI (currently in alpha) to become stable sooner and be the best option for debugging, please do submit feedback and contribute if you can.
EDIT: we now have the Visual Studio Code IDE support with first-class debug support: https://github.com/intuit/karate/wiki/IDE-Support#vs-code-karate-plugin
EDIt2: If you want to debug Java code, that is possible with the new IntelliJ plugin: https://plugins.jetbrains.com/plugin/19232-karate
As per the documentation here, At this moment best way to debug Karate Steps is using Visual Studio Code for developing tests and VS Code Karate Plugin for debugging.
Visual Studio Code is Free, built on open Source and runs on all platforms including mac/linux and windows.
Please note this
The Karate UI has been retired and is not available in 0.9.5 onwards !
Use the VS Code Debug Support instead.
As per the comment by Peter Thomas, Eclipse/IntelliJ may also support debugging but I am unable to find any development there.