<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Furkan on Medium]]></title>
        <description><![CDATA[Stories by Furkan on Medium]]></description>
        <link>https://medium.com/@furkanyyavuz?source=rss-ecbe5017bdfe------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*qktrVG8w0l5PwWlfzURT1w.png</url>
            <title>Stories by Furkan on Medium</title>
            <link>https://medium.com/@furkanyyavuz?source=rss-ecbe5017bdfe------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 13:35:18 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@furkanyyavuz/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Test Automation Journey: Before the First Milestone]]></title>
            <link>https://furkanyyavuz.medium.com/test-automation-journey-before-the-first-milestone-10ceadeea8b0?source=rss-ecbe5017bdfe------2</link>
            <guid isPermaLink="false">https://medium.com/p/10ceadeea8b0</guid>
            <category><![CDATA[qa]]></category>
            <category><![CDATA[test-automation]]></category>
            <category><![CDATA[selenium]]></category>
            <category><![CDATA[bdd]]></category>
            <category><![CDATA[cucumber]]></category>
            <dc:creator><![CDATA[Furkan]]></dc:creator>
            <pubDate>Mon, 08 Jan 2024 07:40:11 GMT</pubDate>
            <atom:updated>2024-01-08T07:40:11.255Z</atom:updated>
            <content:encoded><![CDATA[<p>Test automation became a topic that is considered essential after software projects started to follow agile principles. Actually, it is a reflection of the software world of fast-paced changing technology. In various fields, from health to agriculture, we are witnessing and hearing about automation solutions. In the software world, we automate our tasks to deliver high-quality products. As explained in the <a href="https://agiletester.ca/agile-testing/">Agile Testing</a> book, which I can call one of the key sources for Software Quality Engineers:</p><blockquote>Good-enough automation frees the team to deliver high-quality code frequently.</blockquote><p>Automating tests brings many benefits to QA Engineers and maybe, the most important one is the safety net. We get feedback if any part of the product covered by automated tests is broken, and we can take prompt action to solve the issue(s). If you want to read more about why test automation is important or what should be the starting steps for it, please check my colleague <a href="https://medium.com/u/c5a4edacfac2">Semih Aydin</a>’s <a href="https://medium.com/@semihaydin/building-test-automation-from-scratch-365b65b3a58b">post</a>. In this article, I will explain the test automation journey at Recognyte by discussing the technology preferences we employed for test automation implementation and the challenges we have overcome.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XYm1T61-XDc5awPAxSYwGQ.png" /></figure><h3>Let’s Start</h3><p>At <a href="https://recognyte.com/">Recognyte</a>, our test automation journey started when we needed to make sure that the main functionalities of our products were working. We know what we want to achieve and set two goals as the first milestone for this test automation journey:</p><blockquote>1. Check main functionalities as automated in the production environment every day</blockquote><blockquote>2. Check main functionalities on the pre-production environment before each release testing</blockquote><p>The motivation behind the first goal is to be sure that there are no major issues in the production environment where we have live clients. If one of the main functionalities is not working for some reason, it needs to be identified as early as possible before the clients have difficulty and is solved.</p><p>The motivation behind the second goal is to be sure that the main functionalities are working before each release testing is initiated. Spending time on a version that has broken the main features and components does not make sense. A version should go into release testing with a bare minimum of core functionalities working as expected. Otherwise, the product is not viable and simply not ready to be tested any further.</p><p>To be able to reach this milestone we decided to build test automation by starting with an end-to-end scenario that covers the main features of the product.</p><figure><img alt="Cucumber, https://cucumber.io/" src="https://cdn-images-1.medium.com/max/261/1*2qL3B5kAuyXiroFbqOukPQ.png" /><figcaption>Cucumber, <a href="https://cucumber.io/">https://cucumber.io/</a></figcaption></figure><p>As a first step, We prepared a feature file that includes <a href="https://dannorth.net/introducing-bdd/">BDD</a> scenarios covering the main features of the product. BDD is one of the key agile practices we follow which brings common understanding to engineering and product teams. This feature file was shared and finalized in collaboration with our business analyst. After we decided to carry on with BDD and completed the feature file, we need to have a framework which facilitates BDD. <a href="https://cucumber.io/">Cucumber</a> is one of the most famous BDD frameworks. If we need to mention about what Cucumber does briefly:</p><blockquote>Cucumber reads executable specifications written in plain text and validates that the software does what those specifications say.</blockquote><p>For more detailed information, you can visit the official Cucumber <a href="https://cucumber.io/docs/guides/overview/">documentation</a> and we also recommend <strong>The Cucumber Book</strong> written by Matt Wynne and Aslak Hellesøy.</p><h4>Selenium WebDriver</h4><p><a href="https://www.selenium.dev/">Selenium</a> is probably the first thing that you will see after googling “web application automation testing”. It’s a free widely used open source tool with a great community. Briefly, it enables you to write code for controlling and interacting with web browsers, executing actions such as clicking buttons, completing forms, navigating web pages, and confirming web page content. By using Selenium WebDriver API and necessary drivers, we can make the interaction mentioned above.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q1uR6JX6XpGK4lgcUfLN1Q.png" /></figure><p>So, as a second step, we decided to combine <strong>Selenium Java</strong> with Cucumber because our QA team already has the necessary experience with Selenium Java to achieve the first milestone.</p><h4>Selenium Manager or WebDriver Manager?</h4><p>In the early days of Selenium, it was a real pain to manage this process which involves downloading the drivers, checking their compatibility with the browser, and ensuring proper maintenance over time. To manage this process in a fully automated manner, Boni García developed an open-source library <a href="https://bonigarcia.dev/webdrivermanager/">WebDriver Manager</a> which I have been using for a long time.</p><p>However, Selenium already introduced <a href="https://www.selenium.dev/documentation/selenium_manager/">Selenium Manger</a> at <a href="https://www.selenium.dev/blog/2022/introducing-selenium-manager/">4th of November, 2022</a> by saying:</p><blockquote>In addition, a <a href="https://www.selenium.dev/blog/2021/selenium-survey-results/">survey</a> done on January 2021 showed that most Selenium users want to get rid of the driver management problem. Plus, the fact that the <a href="https://www.selenium.dev/documentation/webdriver/troubleshooting/errors/driver_location/">driver installation</a> page is by far the most visited one in the Selenium documentation.</blockquote><p>Whilst there is an alternative for WebDriver Manager which is provided by Selenium itself, why did we prefer to go with WebDriver Manager? The answer is <a href="https://bonigarcia.dev/webdrivermanager/#browsers-in-docker">self-managed browsers in Docker containers</a>. We preferred to run our tests in docker containers because it ensures consistent and portable test environments, making it easy to manage dependencies, and enhancing the efficiency and reliability of the test automation. The easiest solution was provided by WebDriver Manager.</p><p>The <em>webdriver </em>is initiated<em> </em>by using the code below.</p><pre>WebDriverManager wdm = WebDriverManager.chromedriver().capabilities(chromeOptions).browserInDocker().enableVnc().enableRecording();<br>wdm.create();</pre><p>WebDriverManager creates 3 docker containers:</p><ul><li>for chrome browser</li><li>for remote access to the container</li><li>for screen recording</li></ul><p>Our product has an import files feature which should be covered in test automation. For this objective, we must transfer some files that will be employed in scenarios to the Docker containers and we utilized the WebDriver Manager method below: <em>wdm.dockerVolumes()</em></p><p>Interestingly, it was working on my local <strong>with and without</strong> <em>dockerVolume()</em>. However, it is NOT working on the virtual machine (VM) <strong>with</strong> <em>dockerVolume()</em>. Unfortunately, we could not find the cause of this <a href="https://github.com/bonigarcia/webdrivermanager/issues/996">issue</a> and solve it. Then, we did not want to spend time on it and developed a<a href="https://stackoverflow.com/a/77215265/10464527"> side way</a>.</p><h4>Infrastructure</h4><p>As the third step, we started to implement scenarios; selected the TestNG testing framework, and adopted the <a href="https://www.selenium.dev/documentation/test_practices/encouraged/page_object_models/">page-object model</a>. After a while, when we were closer to the first milestone, the DevOps team created a VM on the Google Cloud Platform (GCP), and connected our Github project with the VM. That gave us an opportunity to run our implementations on GCP where our development team manages various operations.</p><h4>Dealing with different environments</h4><p>We have a production environment where we have live clients and a pre-production environment where we perform QA processes. After we completed automating all prioritized scenarios, we had to take the last step:</p><ul><li>including the test automation in CI/CD for <em>pre-production</em> environment</li><li>putting the test automation in a schedule on <em>production</em> environment as working on a daily basis.</li></ul><p>The DevOps team needs to trigger the test automation for the different purposes above and at different times, which means it cannot be triggered with the same command. Because, if a command calls the test automation, how can we understand that the call is for which environment?</p><p>For this reason, we needed to provide them the two command lines that could trigger the test automation for both environments as automated in the DevOps mechanism. On the other hand, we wanted to run completely the same feature file for both environments by logging in different environments. Because duplicating whole the feature file does not make sense. Duplicated files would have the same scenarios except the login scenario. So, we handled this part by using <a href="https://cucumber.io/docs/cucumber/api/?lang=java#tags">Cucumber Tags</a> which help organise your features and scenarios.</p><p>The login scenario was like below:</p><pre><strong>Scenario</strong>: Display DataScout Landing Page<br>    <strong>Given</strong> the user logins with custom token<br>    <strong>Then</strong> the DataScout Landing page is displayed</pre><p>We duplicated it and differentiated by using tags for <em>pre-production</em> and <em>production</em> environments:</p><pre><strong>@preprod</strong><br><strong>Scenario:</strong> Display DataScout Landing Page<br>    Given the user logins with custom token to &quot;preprod&quot; environment<br>    Then the DataScout Landing page is displayed<br><strong>@prod</strong><br><strong>Scenario:</strong> Display DataScout Landing Page<br>    Given the user logins with custom token to &quot;prod&quot; environment<br>    Then the DataScout Landing page is displayed</pre><pre><strong>@preprod</strong><br><strong>@prod<br>Scenario:</strong> first scenario<br>    Given ...<br>    When ...<br>    Then ...</pre><p>As you notice, in the first two scenarios, the <em>Given </em>sentences are same with an input variable (<em>“”</em>) which we use to make conditional configuration in the background. All other scenarios have both tags because we want to run them on both environments.</p><p>Then, we also duplicated the test runner class and they included different <em>tags</em> in <a href="https://cucumber.io/docs/cucumber/api/?lang=java#options">CucumberOptions</a>.</p><p>Test runner file for the <strong>pre-production</strong> environment:</p><pre>@CucumberOptions(<br>        plugin = {&quot;pretty&quot;,<br>                &quot;html:target/cucumber-reports/cucumber.html&quot;,<br>                &quot;json:target/cucumber-reports/cucumber.json&quot;<br>        },<br>        features = &quot;feature-file-directory&quot;,<br>        glue = {&quot;step-definitions-file&quot;},<br>        tags = &quot;@preprod&quot;<br>)</pre><p>Test runner file for the <strong>production</strong> environment:</p><pre>@CucumberOptions(<br>        plugin = {&quot;pretty&quot;,<br>                &quot;html:target/cucumber-reports/cucumber.html&quot;,<br>                &quot;json:target/cucumber-reports/cucumber.json&quot;<br>        },<br>        features = &quot;feature-file-directory&quot;,<br>        glue = {&quot;step-definitions-file&quot;},<br>        tags = &quot;@prod&quot;<br>)</pre><p>Finally, we give the commands below to DevOps for trigger test automation on different environment. The commands refer to different test runner files.</p><pre>//For Prod:<br>mvn clean<br>mvn -Dtest=ProdTestRunner test</pre><pre>//For Pre-Prod:<br>mvn clean<br>mvn -Dtest=PreProdTestRunner test</pre><p>Now, every day at 3 AM, the test automation runs, and results are reported on a website as two files: screen recording of the test run and cucumber results report (like <a href="https://reports.cucumber.io/">this</a>). Now, when we achieve the first milestone, anyone who reads these results can easily understand the test scenarios are or which part of the product failed.</p><p>A test automation project is not something — to implement and leave. It lives as long as the product lives: When the product has a change, the test automation needs maintenance. We keep working on it to expand our test automation coverage.</p><p><strong>Furkan Yavuz</strong> has been a software quality assurance engineer for over 7 years. He holds a master’s degree in information security. Furkan joined Recognyte in 2020 and has been working as QA Lead.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=10ceadeea8b0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[SonarCloud from Scratch]]></title>
            <link>https://furkanyyavuz.medium.com/sonarcloud-from-scratch-43cf8fdf1cd9?source=rss-ecbe5017bdfe------2</link>
            <guid isPermaLink="false">https://medium.com/p/43cf8fdf1cd9</guid>
            <category><![CDATA[clean-code]]></category>
            <category><![CDATA[code-quality]]></category>
            <category><![CDATA[proptech]]></category>
            <category><![CDATA[sonarcloud]]></category>
            <category><![CDATA[qa]]></category>
            <dc:creator><![CDATA[Furkan]]></dc:creator>
            <pubDate>Thu, 24 Aug 2023 12:41:28 GMT</pubDate>
            <atom:updated>2023-08-24T13:28:56.805Z</atom:updated>
            <content:encoded><![CDATA[<p>Code is the most critical asset of tech companies. We build our applications by writing code, but more is needed. We need to shape it well. The term <em>code quality</em> appears at this stage. It briefly indicates the value of the code. There are lots of tools that help to improve products’ code quality, but SonarCloud (from Sonar) is a proven product. In this article, we tell you about our SonarCloud journey from scratch in Recognyte.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dL3I-UgrJy9P9Ozf8LDy0g.png" /></figure><p><a href="https://www.sonarsource.com/products/sonarcloud/">SonarCloud</a> is a code quality tool, a SaaS (Software as a Service) version of Sonar which is derived from the self-managed SonarQube product. As code is the most important asset, companies need to have reliable, maintainable, and secure code. SonarCloud runs static code analysis, gives users the ability to monitor code quality, and helps build more reliable, maintainable, and secure code repositories. It supports many coding languages. The analysis includes unit test coverage and detects vulnerabilities (security breaches), bugs, duplications, and code smells. Code smells are an indicator that says the code needs refactoring to improve performance, maintainability, and readability.</p><p>Before SonarCloud, we used to write unit tests and we could measure the coverage for a repository, but due to the number of repositories (100+), we had to build from scratch or find a more general solution. We would fix known bugs or vulnerabilities, but we couldn’t be sure if there were any undetected bugs yet to be discovered. In essence, we were unable to measure the code quality of our projects, lacked a holistic view, and couldn’t make future plans. Additionally, we couldn’t guarantee that we weren’t introducing issues or code vulnerabilities into production within our CI/CD pipeline.</p><p>We were in need of a code quality tool that could provide us with analysis reports for all our new code, enable us to deliver our new code to production at a certain standard, and help us improve the code quality of all of our projects. After extensive research and analysis, we decided to utilize SonarCloud which aligns with our code stack and VCS (Version Control System).</p><p>SonarCloud has 2 types of analysis methods:</p><ul><li>Automatic Analysis: It is automatically triggered when a project is connected to SonarCloud.</li><li>CI-based Analysis: It needs configuration and integration of code analysis and quality checks into the CI (Continuous Integration) process.</li></ul><p>For getting unit test coverage reports, each language needs CI-based analysis and different configurations. Both analysis methods can help detect vulnerabilities, bugs, duplications, and code smells.</p><p>SonarCloud has CI/CD <a href="https://www.sonarsource.com/products/sonarcloud/features/integrations/">integrations</a> with many workflows and enables automated code reviews. When the required configuration of CI-based analysis is made for a project, all analysis results are given in each PR (pull request) in VCS (version control systems) such as GitHub, GitLab, or Bitbucket. SonarCloud runs its analysis according to the Quality Gate parameters and if the new code does not meet the Quality Gate, the SonarCloud analysis fails.</p><p>To mention a little bit about our company, <a href="https://recognyte.com/">Recognyte</a> is a prop-tech company that has combined years of real estate knowledge alongside artificial intelligence and large-scale data collection to ease property management with its SaaS products.</p><p>At Recognyte, we have 3 different products, 4 main languages in our stack, and more than 120 projects and we are using GitHub as VCS.</p><p>When we decided to go with SonarCloud, our team members had a few concerns and we encountered some resistance. The primary concern was the possibility of this tool reducing productivity and slowing down processes during its implementation and ongoing usage.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*igBv2lS_8bQ0VZBM" /><figcaption><a href="https://www.freepik.com/free-vector/website-faq-section-user-help-desk-customer-support-frequently-asked-questions-problem-solution-quiz-game-confused-man-cartoon-character_11667041.htm#query=question%20mark%20illustration&amp;position=1&amp;from_view=keyword&amp;track=ais">image</a> by vectorjuice from freepik.com</figcaption></figure><p>According to my colleague, Kostya:</p><blockquote>Before SonarCloud, we already had “linters” as a part of CI for all our Go repositories (which is around 80% — 90% of our code base). We are using <em>golangci-lint</em> tool for it, which is, essentially, a list of the most popular linters. But, of course, the tool itself doesn’t give us any visibility on what is going on in the product in terms of test coverage and code smells. So, our concerns were around having a tool that provides something that we already have, but also has fewer linters and requires more steps for developers in their daily workflow.</blockquote><blockquote>But after we had a look into more “advanced” configuration in SonarCloud, we found that almost everything could be adjusted for our specific needs, so SonarCloud “Quality Gate’’ feature has not become a bottleneck for us. Sometimes we have to add more exclusions or ignore some rules, but I would say those are exceptions. We spent some time building Recognyte-specific configuration and SonarCloud started to work for us and give us benefits.</blockquote><p>As a first step, our aim was to convince all teams to get rid of all their concerns. We selected three repositories from different stacks as a starting point and completed all the necessary configurations. Teams encountered SonarCloud analysis results with each of their pull requests in the version control system. Without any specific guidance from us, many of our team members began improving their code by examining the analysis results. This was truly motivating. We conducted individual meetings with each team and explained the tool to everyone. Once all teams recognized the benefits of SonarCloud and came on board, we introduced SonarCloud to all our active projects. Since we had a large number of repositories, we opened Jira tickets for each team explaining how the configurations should be done. Additionally, we created a SonarCloud Implementation Board page in Confluence to track this process.</p><p>We also set the Branch Protection rule in Github by adding SonarCloud checks to block the code which gets failed SonarCloud analysis results. This action is the key to protecting our code base from poor code and keeping our code base in shape according to standards that we set in our Quality Gate.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xLdRpki4gpUlvqMO" /><figcaption><a href="https://www.freepik.com/free-vector/product-quality-control-abstract-concept_12085281.htm#query=quality%20check&amp;position=3&amp;from_view=search&amp;track=ais">image</a> by vectorjuice from freepik.com</figcaption></figure><p>Our development teams have become familiar with SonarCloud. We obtain the SonarCloud analysis results with each PR before merging the new code into our main branches. If our developments fail to meet the Quality Gate (in other words, if the SonarCloud analysis is a fail), we understand that the developed code does not adhere to our standards and we bring the code into shape.</p><p>Throughout this process, we occasionally encountered false positive results (which means SonarCloud says there is an issue but there is not actually), but we resolved these issues with the support of the <a href="https://community.sonarsource.com/">Sonar community</a>.</p><p>Now, we have introduced all our projects to SonarCloud. We can observe the size of our projects, the extent to which the code is covered by unit tests, and the presence of weaknesses and errors in the entire codebase. With the analysis results obtained from each PR, we can assess the state of our new code and monitor the improvement trend in our code quality since adopting SonarCloud. By focusing on new code rather than legacy code with branch protection, we continuously <a href="https://www.sonarsource.com/blog/water-leak-changes-the-game-for-technical-debt-management/">fix the leak</a> in our code repository.</p><p>In conclusion, our journey with SonarCloud has been a transformative experience for Recognyte. By adopting this solution as our code quality tool, we have gained valuable insights into the quality of our code repositories and improved our overall code standards. The result has been an increase in code quality, productivity, and a more reliable and secure software development process. We are delighted with the positive impact SonarCloud has had on our projects and code quality and it ensures that our products are reliable for our clients.</p><p><strong>Furkan Yavuz</strong> has been a software quality assurance engineer for over 7 years. He holds a master’s degree in information security. Furkan joined Recognyte in 2020 and has been working as QA Lead.</p><p><a href="https://medium.com/@gregmus?source=user_profile-------------------------------------"><strong>Kostiantyn Osichenko</strong></a> has 6 years of experience in software development and architecture design as PHP and then, Golang engineer. Currently holds the position of Technical Lead at Recognyte.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=43cf8fdf1cd9" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>