parse(source) == transformed

        where:
        source              | transformed
        "text<p>para 2</p>" | "<p>text</p><p>para 2</p>"
        "text<h2>text</h2>" | "<p>text</p><h2>text</h2>"
    }

    def "does not add implicit <p> element for elements with inline content"() {
        expect:
        parse(source) == transformed

        where:
        source                                   | transformed

    launcherJar.from(tasks.jar)
    agentJars.from(agentsClasspath)
    // The trick below is to use the templates from the current code instead of the wrapper. It does not cover the case where the generation logic is updated though.
    unixScriptTemplate.from(layout.projectDirectory.file("../../../platforms/jvm/plugins-application/src/main/resources/org/gradle/api/internal/plugins/unixStartScript.txt"))

        You can use [the template](https://github.com/gradle/gradle-issue-reproducer) with a Gradle GitHub action set up to showcase your problem.
        In the rare cases where this is infeasible, we will also accept a detailed set of instructions.
        You can also use [Gradle Project Replicator](https://github.com/android/project-replicator) to reproduce the structure of your project.
    validations:

     * use instead of the default one.
     */
    private static final String MAVEN_WRAPPER_PROPERTIES_PATH =
            ".mvn/wrapper/maven-wrapper.properties";

    /**
     * Path where the maven-wrapper.jar will be saved to.
     */
    private static final String MAVEN_WRAPPER_JAR_PATH =
            ".mvn/wrapper/maven-wrapper.jar";

    /**

 * </ul>
 *
 * Produces the following:
 * <ul>
 * <li>A docbook book XML file containing the reference.</li>
 * <li>A meta-data file containing information about where the canonical documentation for each class can be found:
 * as dsl doc, javadoc or groovydoc.</li>
 * </ul>
 */
@CacheableTask
abstract class AssembleDslDocTask extends DefaultTask {

# Each automation has an `if` key that defines conditions to run the automation,
# as well as a `run` key with the actions to do. All the conditions need to be true
# for the actions to run (there is an implicit AND relation between
# the conditions on each line).  Each condition is specified as a Jinja expression.  You
# can also use the constant `true` if you want to always run the actions.

# Each automation has an `if` key that defines conditions to run the automation,
# as well as a `run` key with the actions to do. All the conditions need to be true
# for the actions to run (there is an implicit AND relation between
# the conditions on each line).  Each condition is specified as a Jinja expression.  You
# can also use the constant `true` if you want to always run the actions.

     * classloader by {@link JApiCmpWorkerAction}, so there is no way to use a
     * normal property on the renderer instance and just set the location of the API file in it.
     *
     * Instead, we'll encode the path to the file in the description data field, as a link.  This is
     * useful regardless of this renderer's needs, since now there will be a link embedded in the report

# Each automation has an `if` key that defines conditions to run the automation,
# as well as a `run` key with the actions to do. All the conditions need to be true
# for the actions to run (there is an implicit AND relation between
# the conditions on each line).  Each condition is specified as a Jinja expression.  You
# can also use the constant `true` if you want to always run the actions.

    }

    protected boolean isOverride(JApiMethod method) {
        // No point in parsing the source file if the method is not there any more.
        if (method.changeStatus == JApiChangeStatus.REMOVED) {
            return false
        }
        // @Override has source retention - so we need to peek into the sources