It is common practice for unit testing in the Spring JUnit test to load files and their content to make more robust test cases. Or we have a test case where we need perhaps read or write data or test specific files in general.

1. Using ClassLoader

First option is to use and instance of the ClassLoader.

final ClassLoader classLoader = getClass().getClassLoader();
final String resourceName = "daily_report.csv";
final File file = new File(classLoader.getResource(resourceName).getFile());
final String absolutePath = file.getAbsolutePath();

System.out.println(absolutePath);

Assert.assertTrue(absolutePath.endsWith("/daily_report.csv"));

We will get the output of the absolute path as:

/Users/your.user.name/DEV/tutorials/testing-modules/junit-5-basics/target/test-classes/daily_report.csv

Be aware that the path to the resource will look slightly different from the rest of the examples. It is because the ClassLoader looks for the resources on the classpath. Therefore obtaining a path to the resource is from the point of the compiled project.

If you build your project with Maven, compiled classes and resources are in the /target directory in the project root folder. That is why you need to count on the path to resource adjusted for the difference. The other two methods use project root.

2. Using Path

Second option is to use an instance of the Path class. We will call a static factory method Paths.get() which will return us its new instance. Then we will convert Path to File instance and from the that we will extract absolute path with getAbsolutePath() method.

final Path resourcePath = Paths.get("src","test","resources");
final String absolutePath = resourcePath.toFile().getAbsolutePath();

System.out.println(absolutePath);

Assert.assertTrue(absolutePath.endsWith("src/test/resources"));

We will get the same output as in example using java.io.File:

/Users/your.user.name/DEV/tutorials/testing-modules/junit-5-basics/src/test/resources

Path class was introduced in Java 7; consequently, it is available also in older and legacy projects.

3. Using java.io.File

Third approach is to use an instance of the java.io.File class to read the /src/test/resources directory. We will create instance and then we will call its method getAbsolutePath() method. Path we will obtain in this way is relative path towards the current working directory of the project.

final String resourcePath = "src/test/resources";
final File file = new File(resourcePath);
final String absolutePath = file.getAbsolutePath();

System.out.println(absolutePath);

Assert.assertTrue(absolutePath.endsWith("src/test/resources"));

Let’s see an example output when running on the Unix based system:

/Users/your.user.name/DEV/tutorials/testing-modules/junit-5-basics/src/test/resources

Conclusion

This article was a quick tutorial where we showed how to read resources from the /src/test/resources directory in JUnit 5.

There are three different ways how to get a resource for testing. We can either use File, Paths or ClassLoader class.

As always, you can find all our examples on our GitHub project!

Spring Boot application default port

When you develop long enough with Spring Boot, you know the error very well. You work on your Spring Boot application, and you use a console for running the application. Or you use your favourite IDE runner for starting your application (my preferred IDE is IntelliJ IDEA). Your Spring Boot application in the running embedded in servlet container when suddenly you need to do something else. Instead of adequately terminating the embedded servlet container with your application, you detach from the servlet by a closing terminal window or switching to a different project in IDE. However, because you did not correctly terminate your IDE runner and did not correctly kill the terminal’s application process, the servlet with the application is still running. After that, you might, of course, start developing something else. Surprise, surprise, you cannot. Any Spring Boot application port, by default number 8080, is still occupied with the unterminated process of your Spring Boot servlet container. (In the tutorial, we assume you are using the Spring Boot application port on default port number 8080).

I guess you might be a little bit confused by all the wording describing the issue. Let’s take a look first at some terminology to understand the problem here fully:

What is process?

We are aware that readers of this blog are pretty skilled in computers. So, we will provide a brief description of what is computer process without going to further details.

Every time you start a program on your computer, it requires a certain amount of resource. Your Central Processor Unit, known as CPU, add to this program its process.

If you open task manager on your Windows, you will see a list of all process your CPU is trying to manage at once.

What is process ID?

The process ID, also known as PID, is the number assigned by the operating system to each new process.

What is port number?

A port number is how to identify individual processes for network message when they arrive on a server. In a real sense, a port number is the logical address of each application or process that helps identify the sender and receiver of processed messages.

You can imagine port numbers as doors in the hallway into which message flows through the networks. It leaves one door and enters another.

What is Spring Boot server container?

Spring Boot is a Spring project which makes Spring development fast. Without going too much into the Spring Boot project details, your Spring Boot application runs in a server container, better known as a servlet. This servlet runs in one computer process and lister on port 8080 for traffic.

Restart computer

One way how to solve this is to restart your computer. During the restart of Windows OS, start all the fresh, and port 8080 will be empty. But this is not the fastest and definitely not the right way for developers to solve this.

Set different port number

Another way is to solve this issue by assigning the Spring Boot application a different port number. When we set a different port number for the application, we will not kill the existing process with the application servlet container. However, we will create a new-different process for the servlet container, which will be listening on a different port. In this way, port 8080 will be still occupied, but we will at least be able to continue to develop.

Go to Spring Boot application.properties file and set different port number for server.port .

Despite this quick workaround, there is a much more mature way to solve this problem. I would like you to show a better way to kill a Spring Boot servlet container listening on port 8080.

Kill the process listening at port 8080

To kill the process listening at port 8080 in the Windows environment, we need only go through a few steps. First, check the image below to get the idea.

Find the process occupying port 8080

On the Windows, open your CMD terminal. With simple command netstat -ano | findstr PORT_NUMBER, where for PORT_NUMBER add port number 8080, and on command prompt output (hit enter), we will get the process id of process listening on port 8080.

Kill the process with command prompt

Now, all we need to do is just type the following command taskkill /F /PID PROCESS_ID and replace PROCESS_ID by the process id we got from the previous command result.

If you have typed everything correctly, process listening on port 8080 should be terminated.

We have recently looked on how to download file from Azure Cloud Storage in legacy application. With the help of Camel and in Java project, all was written in a couple of minutes. But what if you have a special case and your file is changing periodically. For example, every day or every minute?

This article will show you a couple of code examples on how to initiate an action when periodic change is required. This article will also show you how to start the action with Camel’s Scheduler and Camel’s Quartz2 components, mostly for your legacy Camel projects.

We will extend code from an article about downloading a file from Azure Cloud Storage in our hypothetical scenario. We will create Spring Component in our new application for Camel messaging. It will download a certain file from Azure Cloud Storage periodically.

Using Camel Scheduler component as trigger

First of all, we need to extend RouteBuilder. The RouteBuilder is a base class which is derived from to create routing rules using the DSL. Instances of RouteBuilder are then added to the CamelContext.

In the extended class ScheduleFromAzureCloudStorageToDisk we will override configure() method. configure() method is actually used for Camel’s messaging system. We start with Camel’s from method for building routes and we will place our trigger there. We will use the Scheduler component with combination of delay method option for the file download in this example. This delay will trigger periodical lookup into the Azure Cloud Storage. Delay needs to be set up in milliseconds.

The second, thing to do is to set route id for Camel context. Any unique name for route name will do it.

Finally, all we need to do is create a method that will look into the Azure Cloud Storage and save the new document to local/external disk/source. To get a better idea how this method might look in the legacy application, check the article How to download a file from Azure Claud Storage.

package com.codepills.routes;

import org.apache.camel.Exchange;
import org.apache.camel.builder.RouteBuilder;
import org.apache.log4j.Logger;
import org.springframework.stereotype.Component;

@Component
public class ScheduleFromAzureCloudStorageToDisk extends RouteBuilder {

    protected Logger logger = Logger.getLogger(getClass());

    @Override
    public void configure() throws Exception
    {
        logger.info("Start setting up FromAzureCloudStorageToDisk");

        from("scheduler://fileInAzureBlobStorage?delay=30000")  // Trigger action every 30 seconds, number is in milliseconds
            .routeId("FromAzureStorageToDiskRoute")
            .process(this::downloadFromAzureCloudStorageToDisk); // .process((exchange) -> downloadFromAzureCloudStorageToDisk(exchange))

        logger.info("Finished setting up FromAzureCloudStorageToDisk");
    }

    private void downloadFromAzureCloudStorageToDisk(Exchange exchange)
    {
        ...
    }

}

Using Camel Quartz2 component as trigger

If you are working with a legacy application, you might use older versions of Camel. In older version of Camel, Camel does not have support for pure CRON component.

Instead, it has Quartz and Quartz2 components that provide a scheduled delivery of messages using the Quartz Scheduler version 2.x.

Difference between Quartz and Quartz2 is in its API on lower levels and availability in Camel version. Noticeable distinguish in the fact that Quartz2 offers a few more options. So selection and usage of correct Quartz component I will let decision on you.

package com.codepills.routes;

import org.apache.camel.Exchange;
import org.apache.camel.builder.RouteBuilder;
import org.apache.log4j.Logger;
import org.springframework.stereotype.Component;

@Component
public class ScheduleFromAzureCloudStorageToDisk extends RouteBuilder {

    protected Logger logger = Logger.getLogger(getClass());

    @Override
    public void configure() throws Exception
    {
        logger.info("Start setting up FromAzureCloudStorageToDisk");

        from("quartz2://import/azure-file?cron=0/30+*+*+*+*+?") // CRON format for every 30 seconds
            .routeId("FromAzureStorageToDiskRoute")
            .process(this::downloadFromAzureCloudStorageToDisk); // .process((exchange) -> downloadFromAzureCloudStorageToDisk(exchange))

        logger.info("Finished setting up FromAzureCloudStorageToDisk");
    }

    private void downloadFromAzureCloudStorageToDisk(Exchange exchange)
    {
        ...
    }

}

Let’s explain code above. We have created Spring component ScheduleFromAzureCloudStorageToDisk extending RouteBuilder class. We will implement trigger in configure() method. We will place Quartz2 component into the from method, which will build a route for Camel. Route in from method will have form as quartz2 component, then import trigger, then unique id name of import trigger (in our case azure-file) and lastly define time trigger in CRON format after cron parameter for import Quartz2 trigger.

Again, regarding downloadFromAzureCloudStorageToDisk(final Exchange exchange) method, you can find its implementation in article How to download a file from Azure Claud Storage .

That is all. All magic regarding triggering on specific time happens in CRON setting.

How to set up CRON

Since I know how hard and frustrating it is sometimes to set correct CRON format, here are a couple of examples to distinguish for you the order od time elements in CRON used in Quartz

Examples:

• 0/30+*+*+*+*+? Every 30 seconds
• 0+0+6+*+*+? Every day at 6:00 AM
• 0+*+0/6+*+*+? Every 6 hours

If you would like to know how to write a CRON job, check the article How to write CRON jobs.

Whether you want to build your SpringBoot app into the executable JAR file or build classes for a package, remember to use the single Maven flag parameter mvn install -DskipTests, and in the most cases, you will be fine.

However, the rabbit might run deeper, so it will be better to write a bit more info.

-DskipTests option only works from Surefire 2.4. Unfortunately, there are no error messages if you do try to use it with Surefire 2.3 or lower. Therefore, your first task is to check your version of Surefire. For this, run mvn -X test and look for a line that mentions surefire. It should look like this: maven-surefire-plugin: resolved to version 2.3 from repository central. If you use pre-Surefire-2.4 start using -Dmaven.test.skip=true.

Fortunately beyond 2020 older version usage shrinks and you should remember to always favor a -DskipTests above -Dmaven.test.skip=true.

Surefire itself is the intermediate plugin between Maven and JUnit/TestNG for which while -DskipTests works and -Dmaven.test.skip=true works for compiler.

You can find (not really much more) additional info about shortening the build time trough skipping the project tests at maven skipping tests site.

You can add this plugin configuration to your pom if you do not want to set command line arguments:

<plugin>
    <groupId>org.apache.maven.plugins</groupId>
    <artifactId>maven-surefire-plugin</artifactId>
    <configuration>
        <skipTests>true</skipTests>
    </configuration>
</plugin>

Or you can also set skipping tests in you POM file over property:

<properties>
    <maven.test.skip>true</maven.test.skip>
</properties>

Bonus

You can also skip tests in IntelliJ Idea. Just go to Maven tool tab and hit ‘Skip Test’ mode.

I started investigating the cause of the issue. First, I examined with Spring backend REST endpoint call with the breakpoint. The entity with the date was already coming incorrect from the Angular frontend. The backend endpoint on entry showed incorrect data, so the error was automatically not in the backend code.

At the frontend call for the backend REST endpoint, I also used JSON.stringify(object) to check if the produced entity which was leaving the frontend to the backend was correct. What I found was that it was also already incorrect.

Consequently, the issue happened somewhere between the user browser and Angular call for the backend. In this way, I also eliminated the possibility that the problem was somewhere in the Spring JSON conversion library (I used Jackson).

In the end, I figured it out. The issue was in class, which was migrating datepicker form data into the object. While the Angular calendar datepicker collect the date from the browser also with local time zone as a string, data to JSON ware pushed as a JavaScript Date class without a time zone. So if you, for example, pickup up any date, it was selected as the date for midnight with the timezone offset and transferred from string to Date as the day before due the offset counted as date subtraction.

I was working in Central Europe. In the case of Central Europe (CET), where I was working, the browser records +1 hour from GMT in datepicker based on local PC time settings. So if I pick up the date, hours and minutes are selected for midnight (00:00). However, Date class takes the datepicker string and subtract the negative offset from midnight and set a selected date for a day before.

This is the Angular’s datepicker component HTML code from the component form:

<div class="ui-g-6">
    <div>
        <label for="startDate">Start Date</label>
    </div>
    <vrm-form-field [control]="editForm.controls['startDate']">
        <p-calendar id="startDate"
            [formControlName]="'startDate'"
            dateFormat="yy-mm-dd"
            yearRange="1900:2100">
        </p-calendar>
    </vrm-form-field>
</div>

Following static function created an object from form datepicker which was pushed to REST endpoint before the fix:

static mapFromFormData(form: any) {
    return new ImportantDate(form.startDate);
}

To fix the problem, I created a workaround. I made a Utility class with moment library, which takes a datepicker string and select only required date information.

export class Utils {
  public static formatDate(date: any): string {
    if (date === undefined || date === null) {
      return undefined;
    } else {
      return moment(date).format('YYYY-MM-DD');
    }
  }
}

You can see the implementation of the static Utils method, which fixed the date.

static mapFromFormData(form: any) {
    return new ImportantDate(
        Utils.formatDate(form.startDate)
    );
}

Here is a working example of DocumentRepositoryTest class for reference:

@RunWith(MockitoJUnitRunner.class)
public final class DocumentRepositoryTest {

    @Mock
    private DocumentRepository documentRepository;

    @Test
    public void testConvertToNewDocument() {
        String paperColor = new String("white");
        Mockito.when(this.documentRepository.getPaperColor()).thenReturn(paperColor);
        String color = this.documentRepository.getDocumentColor();
        Assert.assertThat(color, is(new String("white"));
    }
}

DocumentRepositoryTest class is mocking documentRepository object.

So you are running your test and suddenly you see NullPointerException:

java.lang.NullPointerException at com.your.custom.clazz.Method.insertTest(CustomServiceTest.java:50)

You looked it up, and it seems your test is written correctly. So what might be a problem if you know how the mocking works?

If you are sure by your mocking skills, the issue will be probably somewhere else. And most likely very trivial. Here is a list of 3 things you should check out.

1. Return something for your Mock.

Most likely, you mistyped returning function. You probably wanted to return the value for the mocked object. So instead of when-thenReturn , you might type just when-then. Maybe it was IntelliSense. Maybe you did it accidentally. But for sure, NullPointerException happened because you want something which is not there. Debug and check if you are returning something.

2. Specify Mockito running class

Don’t forget to annotate your Testing class with @RunWith(MockitoJUnitRunner.class). Most of the people just forget to specify the test runner, running class for mocking.

3. You need to annotate the mocking object with the @Mock annotation

If you want to mock an object, you need to annotate the object with @Mock annotation.