<?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 Arni Hernandez on Medium]]></title>
        <description><![CDATA[Stories by Arni Hernandez on Medium]]></description>
        <link>https://medium.com/@jorgearni?source=rss-58f20f14b186------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*VYlOoP4wEx-oWI_u.</url>
            <title>Stories by Arni Hernandez on Medium</title>
            <link>https://medium.com/@jorgearni?source=rss-58f20f14b186------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 09:02:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jorgearni/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[Running robot framework scripts in a docker container]]></title>
            <link>https://medium.com/@jorgearni/running-robot-framework-scripts-in-a-docker-container-f049cd0c035a?source=rss-58f20f14b186------2</link>
            <guid isPermaLink="false">https://medium.com/p/f049cd0c035a</guid>
            <category><![CDATA[infrastructure]]></category>
            <category><![CDATA[robot-framework]]></category>
            <category><![CDATA[testing]]></category>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[qa]]></category>
            <dc:creator><![CDATA[Arni Hernandez]]></dc:creator>
            <pubDate>Wed, 14 Feb 2018 03:14:02 GMT</pubDate>
            <atom:updated>2018-02-14T03:14:02.618Z</atom:updated>
            <content:encoded><![CDATA[<p>At Tacit Knowledge, Quality Assurance (QA) process is integrated into the Software Development Lifecycle. QA is a source of feedback at all the development stages.</p><p>We work with at least 3 pre-production environments: Development, QA, UAT/Demo. Then, by running different levels of testing we ensure that everything is working well.</p><p>This requires the use of automated checks/tests to verify vital user paths and critical functionality. Most of our automated functional tests are written in <a href="http://robotframework.org/">Robot Framework</a>. Scripts are automatically started by the CI server after a deployment of new code.</p><p>By running these scripts we minimize the risk of introducing regression bugs. They work as a source of information/feedback providing a quick health check.</p><p>This works perfectly as part of our continuous integration process. But what if I don’t want to break the build when introducing new code? I don’t want to pay for lunch because I broke the build. Run the scripts in your local environment you must!</p><p>There are also scenarios in which we have the classic: <em>it works on my machine</em>or <em>I cannot replicate it.</em> These problems are caused by the different platforms that we use: our continuous integration is running on Linux, we use Macs for our development environment, and sometimes we work with teams that use Windows machines.</p><p>In the past, we manually installed all the prerequisites to run our Robot test suite. This ended up being a tedious process as nobody ever installed the toolkit: <a href="https://www.youtube.com/watch?v=8cT_Ulmcrys">ain’t nobody got time for that</a>. Then, we tried using build tools like <a href="https://maven.apache.org/">Maven</a> and <a href="https://pypi.python.org/pypi/Paver/">Paver</a>: they worked for a while, but they required additional work such as creating/maintaining things like POM files and compiling some libraries using <em>Jython</em>.</p><p>Then, we found <a href="https://www.docker.com/">Docker</a>! After a quick read of the documentation and the completion of the tutorial, we decided to integrate our <em>Robot Framework</em>scripts into <em>Docker</em>. The reasons behind this are very simple:</p><ul><li>Ease the setup process of the whole toolkit.</li><li>Provide a simplified way to share and run our automated test scripts.</li><li>Standardized testing environment. The same VM is running on different platforms.</li></ul><p>We started creating our first <em>Dockerfile</em>:</p><pre>FROM ubuntu:latest<br>RUN apt-get update<br>RUN apt-get install -y python-pip python-dev gcc phantomjs firefox<br>RUN pip install robotframework<br>RUN pip install robotframework-sshlibrary<br>RUN pip install robotframework-selenium2library</pre><p>When we ran this image in a <em>Docker</em> container, we discovered that we needed a display in order to run our scripts in Firefox. So we installed <em>Xvfb</em>. It was required to initialize and demonize a <em>Xvfb</em> virtual display so we had to create a couple of bash scripts and we added them to our <em>Dockerfile</em>:</p><pre>ADD xvfb_init /etc/init.d/xvfd<br>RUN chmod a+x /etc/init.d/xvfb<br>ADD xvfb-daemon-run /usr/bin/xvfb-daemon-run<br>RUN chmod a+x /usr/bin/xvfb-daemon-run</pre><p>Next, we decided to automatically checkout the code, thus we added one more line:</p><pre>RUN git clone ssh://&lt;USER&gt;@&lt;GIT_REPO&gt;/&lt;PATH&gt;</pre><p>We finally had a proper setup for running <em>Robot</em> test scripts. It was time to invoke our automated test suite! So we created another bash script (<em>run_smoke.sh</em>) that contained a <em>pybot</em> invocation. Then we added it to our <em>Dockerfile</em>:</p><pre>ADD run_smoke.sh /storefront-qa/robot/<br>RUN chmod a+x /storefront-qa/robot/run_smoke.sh</pre><p>It’s done!, everything is in place!. If you want to test this solution just run the following command in a terminal:</p><pre>docker run -ti bash -c &quot;./run_smoke.sh&quot;</pre><p>Hope you enjoy!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f049cd0c035a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Avoid object recognition frameworks in your test case automation]]></title>
            <link>https://medium.com/@jorgearni/avoid-object-recognition-frameworks-in-your-test-case-automation-952fe9cdebee?source=rss-58f20f14b186------2</link>
            <guid isPermaLink="false">https://medium.com/p/952fe9cdebee</guid>
            <category><![CDATA[quality-assurance]]></category>
            <category><![CDATA[qa]]></category>
            <category><![CDATA[testing]]></category>
            <category><![CDATA[object-recognition]]></category>
            <dc:creator><![CDATA[Arni Hernandez]]></dc:creator>
            <pubDate>Wed, 14 Feb 2018 03:12:30 GMT</pubDate>
            <atom:updated>2018-02-14T03:12:30.732Z</atom:updated>
            <content:encoded><![CDATA[<p>There are shiny tools that promise to automate functional cases using object recognition technology.</p><p><em>Sikuli</em> and <em>Eggplant</em> are examples of tools that allow you to interact with UI elements using images. It’s easy to fall for the promises of using these tools to automate almost anything: from native apps to complex web sites.</p><p>There are several things to consider before you invest time using these kind of tools.</p><h4>Test Case Creation</h4><p>This task can quickly become overwhelming. These tools require a set of images to use as identifiers to click, fill or interact with any element in your app.</p><p>I dare you to checkin your source code to a control system and see the commit that added 100 new files to the repository just to automate 1 single test case.</p><h4>Test stability</h4><p>Although there are ways to customize image recognition sensitivity on all tools, you will soon discover that even small changes will break your tests. Forcing you to either capture images again (and again and again) or accept the fact that your test will be failing from time to time.</p><h4>Test portability</h4><p>If you are planning to collaborate and create or run Test Cases using computers with different OS (Win, Mac, Linux ) or even people that are using different resolution on their screens, this might easily become a source of stress and frustration. Image recognition is just a comparison between two images and as I already stated even small changes might cause a failure.</p><p>Imagine plugging these cases into a CI server. You will create the perfect job to fail every time somebody is pushing code to the environment.</p><h4>Data driven capabilities</h4><p>When doing data driven testing you sometimes want to randomize your data in order to interact with different elements on a website. For example, you might want to choose a randomly selected set of products on a search landing page; in this case you will need to create a new image for each of the products you might interact with. Maintaining all these images and the scripts that use them quickly becomes cumbersome and difficult to maintain.</p><h4>Test Speed and Computing Resources</h4><p>If you do not mind waiting 10–15 minutes to run ( and see them fail) object recognition is your best choice. Also consider the computer power used to run these kind of tools. Image recognition is expensive when you talk about computer cycles.</p><p>In summary I would say that these kind of technologies might have an application ( Super stable UI, running only in 1 machine, etc.. ). But considering the complexity of the current apps you might want to take a look to other tools that offer you a maintainable and stable set of test cases.</p><p>Some people would say that they would use <em>Sikuli</em> to test how the site renders on a device, and I would agree if we did not have access to more suitable tools like <a href="https://github.com/Huddle/PhantomCSS">PhantomCSS</a> that allows you to create a set of images used as a baseline (using CSS locators ) in the same computer that you will run the comparisons (IOU a blog post about PhantomCSS).</p><p>Thanks for reading. Enjoy your testing!!!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=952fe9cdebee" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Three ways to kill a QA engineer]]></title>
            <link>https://medium.com/@jorgearni/three-ways-to-kill-a-qa-engineer-a45d346bf33b?source=rss-58f20f14b186------2</link>
            <guid isPermaLink="false">https://medium.com/p/a45d346bf33b</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[testing]]></category>
            <dc:creator><![CDATA[Arni Hernandez]]></dc:creator>
            <pubDate>Fri, 15 Jul 2016 13:33:38 GMT</pubDate>
            <atom:updated>2016-07-19T01:35:07.539Z</atom:updated>
            <content:encoded><![CDATA[<p>I have been working as a QA engineer for quite a while and fortunately/unfortunately I have faced several situations in which my team, the project or other circumstances were playing against my team, the customer and myself.</p><p>My colleagues might feel identified with this 3 things, but there are other teammates that might not be familiar with these situations. That’s why I decided to write this blog post.</p><h4>Force him to perform repetitive tasks.</h4><p>Put him to write TC the whole day, tell him to automate all the TCs that nobody automated, force him to create reports to justify his work and them isolated from their teammates. You will have a psychopath after a couple of days.</p><p>A QA engineer needs to feel that he is helping his team doing meaningful things. We are not machines, we are not factory workers, we do not love to do repetitive things all day long ( I think nobody does). And after all you have a QA strategy and a methodology, please stick to it.</p><h4>Blame him when things are not working.</h4><p>“I see this story going back and forth” and Why is QA taking so much time to re-test ? Are really common questions to annoy QAs.</p><p>When somebody is testing and finds that something is not well implemented and takes that story back to development, it is a normal thing, that’s why we have that beautiful Jira workflow. But if this happens with every single story, believe me, the problem is not the QA team.</p><p>Things causing these kind of situations might range from stories that are not clear or concrete or it could be the rookie dev team that doesn’t understand the project or stories.</p><p>If you don’t hear your QA team when pointing to problems, you are going to have a bad time. QA is your source of feedback, easing their job, will ease yours.</p><p>Provide solutions having quality in mind, a frustrated QA team is a problem that you don not want to deal with.</p><h4>Do not respect his job.</h4><p>Can you save time and run this in 2 days instead of 4 ? Can you just test this story it is super urgent ? Can you stop whatever you are doing and do this instead ? Could you please stop automating and verify stories ?</p><p>Doing this once because something unexpected happened is OK. But turning this into a routine is when you need to re-check planning and prioritization.</p><p>Testing requires brain power (concentration, analysis, etc… ). If you are switching priorities every couple of days, you will end up having stories that were tested in episodes meaning: waste of time, bad coverage, and frustrated people.</p><p>Testing is a complicated art. And QAs love to respect, improve and propose enhancements to the delivery process. Please be patient and part of the solution. Sit with your QA and find better a sustainable way to deliver and to attend these “unexpected” situations and make them less unexpected ;).</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a45d346bf33b" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>