I want to get the test coverage of my functional / integration automation test cases written in a different repo.
So developers write development code in different repo and integration test cases are in a different repo.
So the question is, is it possible to get the test coverage when both source code and tests are in different repos?
Related
I have to merge my test automation code base with my application development code base in same repo, i have 2 different pom's- one for development code and other for automation testing code.
Dependencies used in development code are not there in testing code and vice-versa. Currently i have merged both the pom's and able to run automated tests on CI/CD pipeline.
But need to have separate pom's for both of them as it might take more time for dev code to build/deploy on CI/CD pipeline because of dependencies present in testing pom.
Please suggest the efficient way to do this?
I have just started learning about test automation in Selenium and found out that most online tutorials would tell you to run the test suite inside an IDE together with a test framework such as TestNG (with testng.xml) and a build tool such as Maven.
When you are working in a software company and told to build a test framework and run automated tests, I don't believe you actually need to fire up your IDE every time you want to execute your test suite. So, my question is, what is the typical setup a software company follows to 'automate' running your test automation scripts?
Software companies are following agile practices and wanna keep up with industry practices. In real projects, CI & CD are used to continuously integrate, deploy and test the software.
Tests are written by SDET using test automation frameworks. While developing test scripts test developers use IDEs like eclipse. However, tests are executed over Jenkins as a job, after required frequency/event.
For example, after every code deployment, Jenkins can automatically trigger your sanity suite, and run regression bi-weekly.
The process' are automated now-a-days with stakeholders demanding agility.
One can invoke selenium java project from command line via .bat file in Jenkins, or using ant/maven as build tools.
IDEs are seldom used to run tests in real world.
I come from a configuration management background using tools like chef. I have done quite a bit of code testing. Recently, I have been entrusted with the responsibility to work on the CI-CD of the various applications. But I am noticing a culture that’s quite not in line with my ideology.
Let’s say we have 4 environments in the CI-CD pipeline: Dev, Test, Stage & Prod. The dev environment is used by the developers to deploy and test the app before rolling out to the next stage (test). The test env is for the QA team to run their test. And STAGE is another env for 2nd layer of test by the QA folks before the code goes to prod.
Now, does it make sense for the QA or the CI process to run unit test while progressing/after deploying the code to (TEST or STAGE)? I agree that the unit test should be a part of the automated build. But unit test is mainly for the developers, for their code testing. The QA should be focus on the functional testing, load testing etc.. using their frameworks which may be selenium/protractor or LISA. Why should they be focusing on junit or nunits?
From my experience the developers do the unit testing.
Where I currently work, we follow TDD (Test Driven Development) so developers write the unit tests for their code first and then they write the actual implementation. This work goes on the Dev environment.
The QA people then do the acceptance tests and review stuff.
I hope this helps.
Unit tests are short code fragments created by programmers or occasionally by white box testers during the development process. Unit tests are typically written and run by software developers to ensure that code meets its design and behaves as intended.
Hope you got the answer.
Your concern is absolutely right. QA does not necessarily be doing Unit Testing (If that's what you mean). This testing is done by dev while they were writing the code (if they were in TDD environment or even if in BDD).
However if QA has used Junit to write their Selenium automated test cases then they will be using Junit to update and change their automated test cases.
It's little difficult to automate Selenium test cases using Junit whereas using other framework like TestNG is the best option to write automation test cases for Selenium.
I hope this answer your question.
We have seperate integration test project which fires the integration test cases on different modules . At present we do not have unit test cases within each module. We would like to ensure that the integration test covers most of the domain functionality.
Since we have the integration tests in a different project , Sonar always reports the test coverage as zero for the modules under test.
Is there any way to have the test coverage reported on a project , when the actual test is run from a different project
Thanks
You should be able to achieve what you want by reading the Code Coverage documentation page on the wiki. Most notably, you'll be able to use the following sample project to see how it works:
IT JaCoCo Sonar Runner sample project
Basically, you have to run your integration tests first using JaCoCo to generate the coverage report (jacoco.exec) and then you reuse this report during the SonarQube analysis.
I am trying to get bytecode coverage analysis using a code coverage tool (like Emma or Jacoco) after testing with a GUI based functional testing tool (like HP QuickTest Pro or Selenium).
Anyone who has done this could please give me an idea to start this project?
I am doing this now. My approach is to use JaCoCo ant tasks to instrument the binary byte-code files, and use a specific CLASSPATH to execute the instrumented binaries from an ant build.xml from Jenkins.
The reason for doing the code coverage from byte-code is that there is an existing set-up that runs test scripts for a large application using HP QuickTest Pro . I would imagine that the test coverage is in the single digits, but we need an empirical baseline to demonstrate the possible improvements in code coverage from doing unit tests during a build.