/test-yo

MarkLogic unit testing sandbox

Primary LanguageJava

This project shows an example of using the marklogic-unit-test framework to run server-side tests within MarkLogic.

Enabling marklogic-unit-test in an ml-gradle project

Using marklogic-unit-test requires two additions to the build.gradle file, as described below.

First, ml-gradle includes an "mlUnitTest" task, which depends on the marklogic-unit-test-client JAR file. ml-gradle does not include this by default (not every ml-gradle user will use marklogic-unit-test), so it must be added to the buildscript:

buildscript {
  repositories {
    jcenter()
  }
  dependencies {
    classpath "com.marklogic:marklogic-unit-test-client:0.12.0"
  }
}

Next, the marklogic-unit-test framework is depended on and installed as an "mlBundle" dependency (the "mlBundle" configuration is a feature of ml-gradle for depending on packages of MarkLogic modules):

repositories {
  jcenter()
}
  
dependencies {
  mlBundle "com.marklogic:marklogic-unit-test-modules:0.12.0"
}

Running unit tests

With the above additions in place, the "mlUnitTest" task can be run. This task will use the value of mlTestRestPort to determine which MarkLogic app server to connect to - see below for how to customize this.

First, deploy the application:

gradle mlDeploy

This will deploy the application along with the marklogic-unit-test modules.

Then, run the tests:

gradle mlUnitTest

Two tests are run, and one should fail, so you can see what a failed test looks like.

This project includes the Gradle Java plugin, which allows you to run tests under src/test/java. This project includes an example of a JUnit Parameterized test that invokes each marklogic-unit-test module separately - you can try it like this:

gradle test

Again, two tests will run, and one will intentionally fail. The Parameterized test can be run in an IDE as well, allowing you to take advantage of your IDE's support for JUnit tests.

You can also access the marklogic-unit-test REST endpoints directly:

And you can run the original UI test runner by going to:

Configuring which server mlUnitTest connects to

Prior to ml-gradle 3.8.1, the mlUnitTest task will connect to mlTestRestPort if it's set, else mlRestPort.

Starting in release 3.8.1, you can configure which REST API server mlUnitTest will connect to. The mlUnitTest task now exposes a property of type DatabaseClientConfig. You can configure the properties of this object, and mlUnitTest will use it for creating a connection to MarkLogic.

Below is an example - note that you need to configure every property necessary for the type of connection you want, as none of the properties of the DatabaseClientConfig have any default value:

mlUnitTest.databaseClientConfig.host = mlHost
mlUnitTest.databaseClientConfig.port = 8880 // probably a port that differs from mlRestPort and mlTestRestPort
mlUnitTest.databaseClientConfig.username = mlUsername
mlUnitTest.databaseClientConfig.password = mlPassword
// Other properties that can be set
// mlUnitTest.databaseClientConfig.securityContextType
// mlUnitTest.databaseClientConfig.database
// mlUnitTest.databaseClientConfig.sslContext
// mlUnitTest.databaseClientConfig.sslHostnameVerifier
// mlUnitTest.databaseClientConfig.certFile
// mlUnitTest.databaseClientConfig.certPassword 
// mlUnitTest.databaseClientConfig.externalName
// mlUnitTest.databaseClientConfig.trustManager