我一直在尝试使用Reflections库，但在使用它时遇到了一些问题，而且为了简单地获取包上的类，我应该包含太多的jar。

我将在这个重复的问题中发布一个解决方案:如何在包中获得所有类的名称?

答案由sp00m撰写;我添加了一些更正，使其工作:

import java.io.File;
import java.io.IOException;
import java.net.URL;
import java.util.Enumeration;
import java.util.LinkedList;
import java.util.List;

public final class ClassFinder {

    private final static char DOT = '.';
    private final static char SLASH = '/';
    private final static String CLASS_SUFFIX = ".class";
    private final static String BAD_PACKAGE_ERROR = "Unable to get resources from path '%s'. Are you sure the given '%s' package exists?";

    public final static List<Class<?>> find(final String scannedPackage) {
        final ClassLoader classLoader = Thread.currentThread().getContextClassLoader();
        final String scannedPath = scannedPackage.replace(DOT, SLASH);
        final Enumeration<URL> resources;
        try {
            resources = classLoader.getResources(scannedPath);
        } catch (IOException e) {
            throw new IllegalArgumentException(String.format(BAD_PACKAGE_ERROR, scannedPath, scannedPackage), e);
        }
        final List<Class<?>> classes = new LinkedList<Class<?>>();
        while (resources.hasMoreElements()) {
            final File file = new File(resources.nextElement().getFile());
            classes.addAll(find(file, scannedPackage));
        }
        return classes;
    }

    private final static List<Class<?>> find(final File file, final String scannedPackage) {
        final List<Class<?>> classes = new LinkedList<Class<?>>();
        if (file.isDirectory()) {
            for (File nestedFile : file.listFiles()) {
                classes.addAll(find(nestedFile, scannedPackage));
            }
        //File names with the $1, $2 holds the anonymous inner classes, we are not interested on them. 
        } else if (file.getName().endsWith(CLASS_SUFFIX) && !file.getName().contains("$")) {

            final int beginIndex = 0;
            final int endIndex = file.getName().length() - CLASS_SUFFIX.length();
            final String className = file.getName().substring(beginIndex, endIndex);
            try {
                final String resource = scannedPackage + DOT + className;
                classes.add(Class.forName(resource));
            } catch (ClassNotFoundException ignore) {
            }
        }
        return classes;
    }

}

要使用它，只需调用本例中提到的sp00n的find方法: 如果需要的话，我还添加了类实例的创建。

List<Class<?>> classes = ClassFinder.find("com.package");

ExcelReporting excelReporting;
for (Class<?> aClass : classes) {
    Constructor constructor = aClass.getConstructor();
    //Create an object of the class type
    constructor.newInstance();
    //...
}

基于@Staale的回答，为了不依赖第三方库，我将通过检查第一个包的物理位置来实现文件系统方法:

import java.io.File;
import java.io.FileFilter;
import java.util.ArrayList;
...
Class<?>[] foundClasses = new Class<?>[0];
final ArrayList<Class<?>> foundClassesDyn = new ArrayList<Class<?>>();

new java.io.File(
    klass.getResource(
        "/" + curPackage.replace( "." , "/")
    ).getFile()
).listFiles(
    new java.io.FileFilter() {
        public boolean accept(java.io.File file) {
            final String classExtension = ".class";

            if ( file.isFile()
                && file.getName().endsWith(classExtension)
                // avoid inner classes
                && ! file.getName().contains("$") )
            {
                try {
                    String className = file.getName();
                    className = className.substring(0, className.length() - classExtension.length());
                    foundClassesDyn.add( Class.forName( curPackage + "." + className ) );
                } catch (ClassNotFoundException e) {
                    e.printStackTrace(System.out);
                }
            }

            return false;
        }
    }
);

foundClasses = foundClassesDyn.toArray(foundClasses);

这是不可能的，因为包中的所有类都可能不会被加载，而你总是知道一个类的包。

在包测试中定义要扫描的类

package test;

public class A {
  private class B {}

  enum C {}

  record D() {}
}

对于org.reflections:reflections:0.10.2，它为我工作如下:

使用反射库扫描包测试中的类

@Test
void t() {
  final String packagePath = "test";
  final Reflections reflections =
    new Reflections(packagePath, Scanners.SubTypes.filterResultsBy(v -> true));
  reflections.getAll(Scanners.SubTypes).forEach(System.out::println);
}

输出

java.lang.constant.Constable
java.lang.Enum
java.lang.Comparable
java.lang.Record
java.lang.Object
java.io.Serializable
test.A$C
test.A$D
test.A$B
test.A

对于io.github.classgraph:classgraph:4.8.146，它为我工作如下:

@Test
void t() {
  final String packagePath = "test";
  try (ScanResult scanResult = new ClassGraph()
    .enableClassInfo()
    .ignoreClassVisibility()
    .acceptPackages(packagePath)
    .scan()) {
    
    scanResult.getAllClasses()
      .forEach(v -> {
        System.out.println(v.getName());
      });
  }
}

输出

test.A
test.A$B
test.A$C
test.A$D

你需要查找类路径中的每个类装入器条目:

    String pkg = "org/apache/commons/lang";
    ClassLoader cl = ClassLoader.getSystemClassLoader();
    URL[] urls = ((URLClassLoader) cl).getURLs();
    for (URL url : urls) {
        System.out.println(url.getFile());
        File jar = new File(url.getFile());
        // ....
    }   

如果条目是目录，只需在右边的子目录中查找:

if (jar.isDirectory()) {
    File subdir = new File(jar, pkg);
    if (!subdir.exists())
        continue;
    File[] files = subdir.listFiles();
    for (File file : files) {
        if (!file.isFile())
            continue;
        if (file.getName().endsWith(".class"))
            System.out.println("Found class: "
                    + file.getName().substring(0,
                            file.getName().length() - 6));
    }
}   

如果条目是文件，并且它是jar，检查它的ZIP条目:

else {
    // try to open as ZIP
    try {
        ZipFile zip = new ZipFile(jar);
        for (Enumeration<? extends ZipEntry> entries = zip
                .entries(); entries.hasMoreElements();) {
            ZipEntry entry = entries.nextElement();
            String name = entry.getName();
            if (!name.startsWith(pkg))
                continue;
            name = name.substring(pkg.length() + 1);
            if (name.indexOf('/') < 0 && name.endsWith(".class"))
                System.out.println("Found class: "
                        + name.substring(0, name.length() - 6));
        }
    } catch (ZipException e) {
        System.out.println("Not a ZIP: " + e.getMessage());
    } catch (IOException e) {
        System.err.println(e.getMessage());
    }
}

现在，一旦包中有了所有的类名，就可以尝试用反射加载它们，并分析它们是类还是接口等等。

谷歌Guava 14包含了一个新类ClassPath，它有三个方法来扫描顶级类:

getTopLevelClasses () getTopLevelClasses(管柱packageName) getTopLevelClassesRecursive(管柱packageName)

有关更多信息，请参阅ClassPath javadocs。