Groovy & Grails: Testing and Deploying

Testing Is Built In

Grails generates a test for every domain class, controller, and service you create. You can write unit tests that run fast without a database, integration tests that include the full application context, and functional tests that drive the real server with an HTTP client.

A Unit Test in Spock

Tests are written in Spock, a Groovy testing framework that reads like plain sentences.

class BookSpec extends Specification {
    def "title cannot be blank"() {
        given:
        def book = new Book(title: "")

        expect:
        !book.validate()
    }
}

The method name is a sentence, and the blocks given, expect, and when organize the test. Here an empty title fails validation, so the test passes.

Running the Suite

grails test

grails test-app

Run the suite often. Compilation errors and test failures are far cheaper to fix now than after deploy.

Packaging for Production

A production Grails app ships as a runnable archive. Grails builds one with a single command.

grails war

The result is a war file in the build folder. A war contains the compiled application and all its dependencies, ready for any servlet container such as Tomcat.

Deploying

Deployment options include dropping the war into a servlet container, running it as a standalone jar, or pushing it to a cloud platform. In production, set the data source to a real database, use environment-specific configuration, and never expose the development console.

• Use SSL and keep secrets in environment variables.
• Set the external database credentials in the production config.
• Run tests in a build pipeline before every release.
• Watch memory and logs, since Grails apps default to a larger heap.

With a green test suite and a war artifact, you are free to deploy wherever the JVM runs.

Key Points

• Grails scaffolds unit, integration, and functional test skeletons.
• Spock specs read like sentences and organize with given, when, and expect.
• grails war packages the application for any servlet container.
• Production config must use a real data source and secure secrets.