* plus is placed in a defer call so the locking is effectively
    * nonexistent.  This patch adds an anonymous function call inside
    * the loop to correctly handle acquiring the write lock with deferred
    * unlock to run at that function's completion, following the normal
    * golang lock/deferred-unlock pattern.

However, this would limit the ability of the Gradle ecosystem plugin authors to take advantage of new Java features and improvements.

One example is when one of the most popular plugins started requiring Java 17 because the underlying framework behind the plugin began requiring 17, and code from the framework was used in the plugin itself.
What would have happened if Gradle was limited to _only Java 11_ at the time?

## Decision

    @Override
    public TaskProvider<? extends Task> createTask(Project project) {
        Configuration loggerUsageConfig = project.getConfigurations().create("loggerUsagePlugin");
        // this makes it easier to test by not requiring this project to be always available in our
        // test sample projects
        if (project.findProject(":test:logger-usage") != null) {

import org.gradle.api.GradleException;

import java.util.List;
import java.util.Optional;

public class JavaUtil {

    /** A convenience method for getting java home for a version of java and requiring that version for the given task to execute */
    public static String getJavaHome(final int version) {
        List<JavaHome> javaHomes = BuildParams.getJavaVersions();

            param("env.JDK8", javaHome(OpenJdk8, Os.LINUX))
            param("env.JDK11", javaHome(OpenJdk11, Os.LINUX))
            param("env.JDK17", javaHome(OpenJdk17, Os.LINUX))
            // Keep 21 and 25 for tests requiring it specifically: https://github.com/gradle/gradle/pull/35410#discussion_r2460835304
            param("env.JDK21", javaHome(OpenJdk21, Os.LINUX))
            param("env.JDK25", javaHome(OpenJdk25, Os.LINUX))

import com.google.common.annotations.GwtCompatible;
import org.jspecify.annotations.NullMarked;
import org.jspecify.annotations.Nullable;

/**
 * To be implemented by test generators that can produce test subjects without requiring any
 * parameters.
 *
 * @param <T> the type created by this generator.
 * @author George van den Driessche
 */
@GwtCompatible
@NullMarked
public interface TestSubjectGenerator<T extends @Nullable Object> {

import com.google.common.annotations.GwtCompatible;
import org.jspecify.annotations.NullMarked;
import org.jspecify.annotations.Nullable;

/**
 * To be implemented by test generators that can produce test subjects without requiring any
 * parameters.
 *
 * @param <T> the type created by this generator.
 * @author George van den Driessche
 */
@GwtCompatible
@NullMarked
public interface TestSubjectGenerator<T extends @Nullable Object> {

# Helpful since the internal CI utilizes a T:\ drive, part of which is mounted
# to the container, and would result in C:\<path> != T:\<path> mismatches,
# when using variables like `TFCI_OUTPUT_DIR` in `docker exec commands,
# requiring conditional path adjustments throughout the CI scripts.
# Note: This does not work for `docker cp` commands.
# For a little more detail, see:

        </field>
        <field>
          <name>requiresDependencyResolution</name>
          <version>1.0.0/1.1.0</version>
          <type>String</type>
          <description>
            Flags this Mojo as requiring the dependencies in the specified class path to be resolved before it can
            execute: {@code compile}, {@code runtime}, {@code test},

     * if this source represents a resolved artifact with known coordinates.
     * <p>
     * This method is primarily used by resolved sources to provide the model ID
     * without requiring the XML to be parsed. For build sources, this typically
     * returns {@code null} since the coordinates are determined by parsing the POM.
     *