We Keep Coding

Java JavaCompiler.run() compiling anonymous classes as well

I am trying to load in text files on the fly and compile them.
File file = new File("Files/"+fileName+".java");
JavaCompiler compiler = ToolProvider.getSystemJavaCompiler();
compiler.run(null, null, errStream, file.getAbsolutePath());
I then will load the compiled .class files later:
public Class loadStrategyClass(File strategyClassFile) throws IOException
{
FileChannel roChannel = new RandomAccessFile(strategyClassFile, "r").getChannel();
ByteBuffer buffer = roChannel.map(FileChannel.MapMode.READ_ONLY, 0, (int)roChannel.size());
return defineClass(strategyClassFile.getName(), buffer, (ProtectionDomain)null);
}
I am currently running into two issues:
The first is if the .java files I load in contain anonymous classes. It doesn't appear that the JavaCompiler class will compile these.
Exception in thread "main" java.lang.IllegalAccessException: Class Loader.ClassLoader can not access a member of class Files.myname.myclass$1 with modifiers ""
The second:
Is that sometimes I will get errors for NoClassDefFoundError:
Exception in thread "main" java.lang.NoClassDefFoundError: Files/myname/myclass
Despite the fact that other classes will load correctly and the .class file is in that path.

Apparently, your loadStrategyClass is defined within a custom ClassLoader. The problem is that it is not enough to call defineClass once for the class you’re interested in, your class loader must be able to resolve classes on demand, usually by implementing findClass, so the JVM can resolve dependencies, like the inner classes.
You didn’t specify, how you get the strategyClassFile argument for the loadStrategyClass method. Since you ran the compiler without any options, I suppose you simply looked up the file relative to the source file. To resolve other dependencies, the actual root of the class directory needs to be known. It becomes much easier when you define where to store the class files, e.g.
// customize these, if you want, null triggers default behavior
DiagnosticListener<JavaFileObject> diagnosticListener = null;
Locale locale = null;
JavaCompiler c = ToolProvider.getSystemJavaCompiler();
StandardJavaFileManager fm
= c.getStandardFileManager(diagnosticListener, locale, Charset.defaultCharset());
// define where to store compiled class files - use a temporary directory
Path binaryDirectory = Files.createTempDirectory("compile-test");
fm.setLocation(StandardLocation.CLASS_OUTPUT,
Collections.singleton(binaryDirectory.toFile()));
JavaCompiler.CompilationTask task = c.getTask(null, fm,
diagnosticListener, Collections.emptySet(), Collections.emptySet(),
// to make this a stand-alone example, I use embedded source code
Collections.singleton(new SimpleJavaFileObject(
URI.create("string:///Class1.java"), Kind.SOURCE) {
public CharSequence getCharContent(boolean ignoreEncodingErrors) {
return "package test;\npublic class Class1 { public class Inner {} }";
}
}));
if(task.call()) try {
URLClassLoader cl = new URLClassLoader(new URL[]{ binaryDirectory.toUri().toURL() });
Class<?> loadedClass = cl.loadClass("test.Class1");
System.out.println("loaded "+loadedClass);
System.out.println("inner classes: "+Arrays.toString(loadedClass.getClasses()));
} catch(ClassNotFoundException ex) {
ex.printStackTrace();
}
In the example above, we know the root of the class directory, because we have defined it. This allows to simply use the existing URLClassLoader rather than implementing a new type of class loader. Of course, using a custom file manager, we also could use an in-memory storage for rather than a temporary directory.
You may use this API to discover what has been generated, which enables you to use the resulting class without knowing beforehand, which package or inner class declarations exist in the source file you’re going to compile.
public static Class<?> compile(
DiagnosticListener<JavaFileObject> diagnosticListener,
Locale locale, String sourceFile) throws IOException, ClassNotFoundException {
JavaCompiler c = ToolProvider.getSystemJavaCompiler();
StandardJavaFileManager fm
= c.getStandardFileManager(diagnosticListener, locale, Charset.defaultCharset());
// define where to store compiled class files - use a temporary directory
Path binaryDirectory = Files.createTempDirectory("compile-test");
fm.setLocation(StandardLocation.CLASS_OUTPUT,
Collections.singleton(binaryDirectory.toFile()));
JavaCompiler.CompilationTask task = c.getTask(null, fm,
diagnosticListener, Collections.emptySet(), Collections.emptySet(),
fm.getJavaFileObjects(new File(sourceFile)));
if(task.call()) {
Class<?> clazz = null;
URLClassLoader cl = new URLClassLoader(new URL[]{binaryDirectory.toUri().toURL()});
for(JavaFileObject o: fm.list(
StandardLocation.CLASS_OUTPUT, "", Collections.singleton(Kind.CLASS), true)) {
String s = binaryDirectory.toUri().relativize(o.toUri()).toString();
s = s.substring(0, s.length()-6).replace('/', '.');
clazz = cl.loadClass(s);
while(clazz.getDeclaringClass() != null) clazz = clazz.getDeclaringClass();
if(Modifier.isPublic(clazz.getModifiers())) break;
}
if(clazz != null) return clazz;
throw new ClassNotFoundException(null,
new NoSuchElementException("no top level class generated"));
}
throw new ClassNotFoundException(null,
new NoSuchElementException("compilation failed"));
}
If you use this to dynamically bind plugins or modules, you may extend the search to look for a result class which implements a particular interface or has a certain annotation.

Parent-Child Classloader Class Resolution

The tasks and initial investigation
I try to set up two Oracle Coherence near cache instances at one java swing application. The idea a solution could be found here. My case is a bit more complicated and this is where the game starts.
Short description
In my case there is an account service. It can have two endpoints: SIT and UAT. In order to create two such services, I need to load two 'instances' of the Coherence in order to override the endpoints with system variables (tangosol.coherence.cacheconfig).
I have:
the main code of the app is located in the mainapp.jar;
the AccountService interface that is located in the account-interfaces.jar;
the AccountServiceImpl class that is located in the account-impl.jar and implements the AccountService interface;
my main application has the following structure
bin: startup.bat, startup.sh
conf: app.properties
lib: mainapp.jar, account-interfaces.jar, account-impl.jar, coherence.jar
Approach tried
I created a dedicated child-first classLoader - InverseClassLoader and made the AppLaunchClassLoader (the default Thread.currentThread().GetContextClassLoader() classLoader) it's parent. With the InverseClassLoader I load the AccountServiceImpl class:
Class<AccountServiceImpl> acImplClass = contextClassLoader.selfLoad(AccountServiceImpl.class).loadClass(AccountServiceImpl.class);
Constructor<AccountServiceImpl> acConstructor =
acImplClass .getConstructor(String.class);
AccountService acService = acConstructor .newInstance(serviceURL);
Issues and questions
I get the 'AccountServiceImpl cannot be cast to AccountService' exceptions, which means that those two classes loaded by different classloaders. But those classloaders are in the parent-child relationship. So am I right that even if a class is loaded by a parent (interface - 'abstract' type) it can't be used with a class (concrete impl) loaded by a child classloader? Why then we need this parent-child relation?
I specified the AccountService interface in a code and it got loaded by a default classloader. I tried wrap the code above is a thread and set the InverseClassLoader it's context classloader. Nothing changed. So am I right that I can't use such interface-implementation coding (as usual coding) and need to use reflection all the time to invoke concrete methods all the time? (Hope there is a solution) ;
Say, I listed both the AccountService and AccountServiceImpl classes for being loaded by the InverseClassLoader. What if I need other classes, that are accessible by those two, to be also loaded by the InverseClassLoader? It there a way to say that all 'related' classes must be loaded by the same classloader?
Update
Here is the InverseClassLoader:
public class InvertedClassLoader extends URLClassLoader {
private final Set<String> classesToNotDelegate = new HashSet<>();
public InvertedClassLoader(URL... urls) {
super(urls, Thread.currentThread().getContextClassLoader());
}
public InvertedClassLoader selfLoad(Class<?> classToNotDelegate) {
classesToNotDelegate.add(classToNotDelegate.getName());
return this;
}
#Override
public Class<?> loadClass(String className, boolean resolve) throws ClassNotFoundException {
if (shouldNotDelegate(className)) {
System.out.println("CHILD LOADER: " + className);
Class<?> clazz = findClass(className);
if (resolve) {
resolveClass(clazz);
}
return clazz;
}
else {
System.out.println("PARENT LOADER: " + className);
return super.loadClass(className, resolve);
}
}
public <T> Class<T> loadClass(Class<? extends T> classToLoad) throws ClassNotFoundException {
final Class<?> clazz = loadClass(classToLoad.getName());
#SuppressWarnings("unchecked")
final Class<T> castedClass = (Class<T>) clazz;
return castedClass;
}
private boolean shouldNotDelegate(String className) {
if (classesToNotDelegate.contains(className) || className.contains("tangosol") ) {
return true;
}
return false;
}

Issue 1, part one I cannot reproduce (see below). As for part 2:
the hierarchy of class-loaders is to prevent the "X cannot be cast to X" exceptions.
But if you break the parent-first rule, you can get into trouble.
About issue 2: setting a thread's context classloader does not do anything in itself, see also this article (javaworld.com)
for some more background. Also, in relation to issue 1, part 2, a quote from the article
that describes what can happen if there is no parent-child relation between the current classloader
and the thread's context classloader:
Remember that the classloader that loads and defines a class is part of the internal JVM's ID for that class.
If the current classloader loads a class X that subsequently executes, say, a JNDI lookup for some data of type Y,
the context loader could load and define Y.
This Y definition will differ from the one by the same name but seen by the current loader.
Enter obscure class cast and loader constraint violation exceptions.
Below is a simple demo-program to show that a cast to an interface from another classloader can work
(note I'm using a simple Java project with classes in a bin-folder and the InvertedClassLoader from your question in the same (test) package):
import java.io.File;
public class ChildFirstClassLoading {
public static void main(String[] args) {
InvertedClassLoader cl = null;
try {
File classesDir = new File(new File("./bin").getCanonicalPath());
System.out.println("Classes dir: " + classesDir);
cl = new InvertedClassLoader(classesDir.toURI().toURL());
cl.selfLoad(CTest.class);
System.out.println("InvertedClassLoader configured.");
new CTest("Test 1").test();
ITest t2 = cl.loadClass(CTest.class)
.getConstructor(String.class)
.newInstance("Test 2");
t2.test();
} catch (Exception e) {
e.printStackTrace();
} finally {
if (cl != null) {
try { cl.close(); } catch (Exception ignored) {}
}
}
}
public interface ITest {
void test();
}
public static class CTest implements ITest {
static {
System.out.println("CTest initialized.");
}
private String s;
public CTest(String s) {
this.s = s;
}
public void test() {
System.out.println(s);
}
}
}
If you change ITest t2 = to CTest t2 = you will get the "CTest cannot be cast to CTest" exception,
but using the interface prevents that exception.
Since this little demo works fine, I'm guessing there is more going on in your application which somehow breaks the class-loading.
I suggest you work from a situation where the class-loading works and keep adding code until it breaks the class-loading.
The InvertedClassLoader looks a lot like the "child first classloader", see this question
for some good answers discussing this manner of class-loading.
The child first classloader can be used to load "related classes" (from your third issue) separately.
You could also update the InvertedClassLoader to always "self-load" classes in certain packages.
And remember that "once a class is loaded by a classloader it uses that classloader to load every other class it needs"
(quote from this blog article).

ClassCastException on custom class loading

I'm trying to write a scripting system in Java and I've managed to get my scripts to compile and instantiate but when I try to cast the script into a "DeftScript" it throws a ClassCastError even thought the script itself extends the class "DeftScript"
Error (the important part at least):
java.lang.ClassCastException: scripts.Compass cannot be cast to com.deft.core.scripts.DeftScript
at com.deft.core.scripts.DeftScriptManager.instantiate(DeftScriptManager.java:52) ~[?:?]
The error is caused by this
deftScript = (DeftScript)obj;
Compiling and Instantiating:
public static DeftScript instantiate(String java) {
File file = new File("./plugins/Deft-Core/scripts/" + java);
DeftScript deftScript = null;
DiagnosticCollector<JavaFileObject> diagnostics = new DiagnosticCollector<JavaFileObject>();
JavaCompiler compiler = ToolProvider.getSystemJavaCompiler();
StandardJavaFileManager fileManager = compiler.getStandardFileManager(diagnostics, null, null);
List<String> optionList = new ArrayList<String>();
optionList.addAll(Arrays.asList("-classpath", System.getProperty("java.class.path") + ";./plugins/Deft-Core.jar"));
Iterable<? extends JavaFileObject> compilationUnit = fileManager.getJavaFileObjectsFromFiles(Arrays.asList(file));
JavaCompiler.CompilationTask task = compiler.getTask(null, fileManager, diagnostics, optionList, null, compilationUnit);
if (task.call()) {
Object obj = null;
try {
String jarFile = "./plugins/Deft-Core.jar";
URLClassLoader classLoader = new URLClassLoader (new URL[] {new File(jarFile).toURI().toURL(), new File("./plugins/Deft-Core/").toURI().toURL()}, Thread.currentThread().getContextClassLoader());
Class<?> loadedClass;
loadedClass = Class.forName("scripts.Compass", false, classLoader);
obj = loadedClass.newInstance();
} catch (ClassNotFoundException | InstantiationException | IllegalAccessException | MalformedURLException e) {
e.printStackTrace();
}
deftScript = (DeftScript)obj;
deftScript.onEnable();
} else {
for (Diagnostic<? extends JavaFileObject> diagnostic : diagnostics.getDiagnostics()) {
System.out.format("Error on line %d in %s%n", diagnostic.getLineNumber(), diagnostic.getSource().toUri());
}
}
return deftScript;
}
Calling the instantiate method:
String script = "Compass.java";
DeftScriptManager.instantiate(script);
DeftScript.java
package com.deft.core.scripts;
public abstract class DeftScript {
public abstract void onEnable();
}
Compass.java
package scripts;
import com.deft.core.scripts.DeftScript;
public class Compass extends DeftScript {
#Override
public void onEnable() {}
}

If your default classloader loads the class DeftScript and the .jar you are loading also contains the class DeftScript, then Java will think that these are two different classes with the same binary name but loaded by different classloaders and you will get that exception since Java sees you trying to mix the two different classes like they are the same thing.
The unique identification of a class in Java consists of the binary class name AND the classloader which was used to load the class.
If you create your URLClassloader like this :
URLClassLoader classLoader =
new URLClassLoader (new URL[] {new File(jarFile).toURI().toURL(),
new File("./plugins/Deft-Core/").toURI().toURL()},
Thread.currentThread().getContextClassLoader());
The second parameter tells java to use the current thread's classloader first to load classes, and only load them from the jar in your URLClassLoader if they are NOT defined in the parent.
Now the classloader will refer to it's parent first and the class DeftScript will only be loaded by the parent classloader even though your .jar file defines the same class (by name).
This was a pretty good article describing the way that this works :
http://www2.sys-con.com/ITSG/virtualcd/java/archives/0808/chaudhri/index.html
This was also helpful
http://www.javaworld.com/article/2077344/core-java/find-a-way-out-of-the-classloader-maze.html?page=1

Reload used classes at runtime Java

I'm working on a program that watches a directory and runs all tests in the directory when it sees changes in the directory.
This requires the program to dynamically load the classes, instead of getting the cached copies.
I can dynamically load the test classes. Changes to the tests get detected and used at runtime. However, this isn't the case for the classes tested by the tests.
My code for dynamically loading the classes and returning a list of test classes:
List<Class<?>> classes = new ArrayList<Class<?>>();
for (File file : classFiles) {
String fullName = file.getPath();
String name = fullName.substring(fullName.indexOf("bin")+4)
.replace('/', '.')
.replace('\\', '.');
name = name.substring(0, name.length() - 6);
tempClass = new DynamicClassLoader(Thread.currentThread().getContextClassLoader()).findClass(name) } catch (ClassNotFoundException e1) {
// TODO Decide how to handle exception
e1.printStackTrace();
}
boolean cHasTestMethods = false;
for(Method method: tempClass.getMethods()){
if(method.isAnnotationPresent(Test.class)){
cHasTestMethods = true;
break;
}
}
if (!Modifier.isAbstract(cachedClass.getModifiers()) && cHasTestMethods) {
classes.add(tempClass);
}
}
return classes;
with DynamicClassLoader being as the Reloader described here How to force Java to reload class upon instantiation?
Any idea how to fix it? I thought all classes would be dynamically loaded. Note however that I don't overwrite loadclass in my DynamicClassLoader because if I do my test classes give init
EDIT:
This doesn't work, the class gets loaded but the tests in it aren't detected...
List<Request> requests = new ArrayList<Request>();
for (File file : classFiles) {
String fullName = file.getPath();
String name = fullName.substring(fullName.indexOf("bin")+4)
.replace('/', '.')
.replace('\\', '.');
name = name.substring(0, name.length() - 6);
Class<?> cachedClass = null;
Class<?> dynamicClass = null;
try {
cachedClass = Class.forName(name);
URL[] urls={ cachedClass.getProtectionDomain().getCodeSource().getLocation() };
ClassLoader delegateParent = cachedClass .getClassLoader().getParent();
URLClassLoader cl = new URLClassLoader(urls, delegateParent) ;
dynamicClass = cl.loadClass(name);
System.out.println(dynamicClass);
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
Edit edit: i detect the test methods like this:
for(Method method: dynamicClass.getMethods()){
if(method.isAnnotationPresent(Test.class)){
requests.add(Request.method(dynamicClass, method.getName()));
}
}

If you used the custom ClassLoader exactly like in the linked answer it is not overriding the method protected Class<?> loadClass(String name, boolean resolve). This implies that when the JVM is resolving dependencies it will still delegate to the parent class loader. And, of course, when it was not delegating to the parent ClassLoader it had the risk of missing some required classes.
The easiest solution is to set up the right parent class loader. You are currently passing Thread.currentThread().getContextClassLoader() which is a bit strange as your main intention is that the delegation should not delegate to that loader but load the changed classes. You have to think about which class loaders exist and which to use and which not. E.g. if the class Foo is within the scope of your current code but you want to (re)load it with the new ClassLoader, Foo.class.getClassLoader().getParent() would be the right delegate parent for the new ClassLoader. Note that it might be null but this doesn’t matter as in this case it would use the bootstrap loader which is the correct parent then.
Note that when you set up the right parent ClassLoader matching your intentions you don’t need that custom ClassLoader anymore. The default implementation (see URLClassLoader) already does the right thing. And with current Java versions it is Closeable making it even more suitable for dynamic loading scenarios.
Here is a simple example of a class reloading:
import java.io.IOException;
import java.net.URL;
import java.net.URLClassLoader;
public class ReloadMyClass
{
public static void main(String[] args)
throws ClassNotFoundException, IOException {
Class<?> myClass=ReloadMyClass.class;
System.out.printf("my class is Class#%x%n", myClass.hashCode());
System.out.println("reloading");
URL[] urls={ myClass.getProtectionDomain().getCodeSource().getLocation() };
ClassLoader delegateParent = myClass.getClassLoader().getParent();
try(URLClassLoader cl=new URLClassLoader(urls, delegateParent)) {
Class<?> reloaded=cl.loadClass(myClass.getName());
System.out.printf("reloaded my class: Class#%x%n", reloaded.hashCode());
System.out.println("Different classes: "+(myClass!=reloaded));
}
}
}

How do I create a parent-last / child-first ClassLoader in Java, or How to override an old Xerces version that was already loaded in the parent CL?

I would like to create a parent-last / child-first class loader, e.g. a class loader that will look for classes in the child class loder first, and only then delegate to it's parent ClassLoader to search for classes.
Clarification:
I know now that to get complete ClassLoading seperation I need to use something like a URLClassLoader passing null as it's parent, thanks to this answer to my previous question
However the current question comes to help me resolve this issue:
My code + dependent jars are being loaded into an existing system, using a ClassLoader that sets that System's ClassLoader as it's parent (URLClassLoader)
That System uses some libraries of a version not compatible with the one I need (e.g. older version of Xerces, that doesn't allow me to run my code)
My code runs perfectly fine if runs stand alone, but it fails if runs from that ClassLoader
Howerver I do need access to many other classes within the parent ClassLoader
Therefore I want to allow me to Override, the parent classloader "jars" with my own: If a class I call is found in the child class loader (e.g. I provided a newer version of Xerces with my own jars, instead of the one users by the ClassLoader that loaded my code and jars.
Here is the System's code that loads my code + Jars (I can't change this one)
File addOnFolder = new File("/addOns");
URL url = addOnFolder.toURL();
URL[] urls = new URL[]{url};
ClassLoader parent = getClass().getClassLoader();
cl = URLClassLoader.newInstance(urls, parent);
Here is "my" code (taken fully from the Flying Sauser "Hello World" code demo):
package flyingsaucerpdf;
import java.io.*;
import com.lowagie.text.DocumentException;
import org.xhtmlrenderer.pdf.ITextRenderer;
public class FirstDoc {
public static void main(String[] args)
throws IOException, DocumentException {
String f = new File("sample.xhtml").getAbsolutePath();
System.out.println(f);
//if(true) return;
String inputFile = "sample.html";
String url = new File(inputFile).toURI().toURL().toString();
String outputFile = "firstdoc.pdf";
OutputStream os = new FileOutputStream(outputFile);
ITextRenderer renderer = new ITextRenderer();
renderer.setDocument(url);
renderer.layout();
renderer.createPDF(os);
os.close();
}
}
This works standalone (running main) but fails with this error when loaded through the parent CL:
org.w3c.dom.DOMException: NAMESPACE_ERR: An attempt is made to
create or change an object in a way
which is incorrect with regard to
namespaces.
probably because the parent system uses Xerces of an older version, and even though I provide the right Xerces jar in the /addOns folder, since it's classes were already loaded and used by the parent System, it doesn't allow my own code to use my own jar due to the direction of the delegation. I hope this makes my question clearer, and I'm sure it has been asked
before. (Perhaps I don't ask the right question)

Today is your lucky day, as I had to solve this exact problem. I warn you though, the innards of class loading are a scary place. Doing this makes me think that the designers of Java never imagined that you might want to have a parent-last classloader.
To use just supply a list of URLs containing classes or jars to be available in the child classloader.
/**
* A parent-last classloader that will try the child classloader first and then the parent.
* This takes a fair bit of doing because java really prefers parent-first.
*
* For those not familiar with class loading trickery, be wary
*/
private static class ParentLastURLClassLoader extends ClassLoader
{
private ChildURLClassLoader childClassLoader;
/**
* This class allows me to call findClass on a classloader
*/
private static class FindClassClassLoader extends ClassLoader
{
public FindClassClassLoader(ClassLoader parent)
{
super(parent);
}
#Override
public Class<?> findClass(String name) throws ClassNotFoundException
{
return super.findClass(name);
}
}
/**
* This class delegates (child then parent) for the findClass method for a URLClassLoader.
* We need this because findClass is protected in URLClassLoader
*/
private static class ChildURLClassLoader extends URLClassLoader
{
private FindClassClassLoader realParent;
public ChildURLClassLoader( URL[] urls, FindClassClassLoader realParent )
{
super(urls, null);
this.realParent = realParent;
}
#Override
public Class<?> findClass(String name) throws ClassNotFoundException
{
try
{
// first try to use the URLClassLoader findClass
return super.findClass(name);
}
catch( ClassNotFoundException e )
{
// if that fails, we ask our real parent classloader to load the class (we give up)
return realParent.loadClass(name);
}
}
}
public ParentLastURLClassLoader(List<URL> classpath)
{
super(Thread.currentThread().getContextClassLoader());
URL[] urls = classpath.toArray(new URL[classpath.size()]);
childClassLoader = new ChildURLClassLoader( urls, new FindClassClassLoader(this.getParent()) );
}
#Override
protected synchronized Class<?> loadClass(String name, boolean resolve) throws ClassNotFoundException
{
try
{
// first we try to find a class inside the child classloader
return childClassLoader.findClass(name);
}
catch( ClassNotFoundException e )
{
// didn't find it, try the parent
return super.loadClass(name, resolve);
}
}
}
EDIT: Sergio and ɹoƃı have pointed out that if you call .loadClass with the same classname, you will get a LinkageError. While this is true, the normal use-case for this classloader is to set it as the thread's classloader Thread.currentThread().setContextClassLoader() or via Class.forName(), and that works as-is.
However, if .loadClass() was needed directly, this code could be added in the ChildURLClassLoader findClass method at the top.
Class<?> loaded = super.findLoadedClass(name);
if( loaded != null )
return loaded;

The following code is what I use. It has the advantage over the other answer that it doesn't break the parent chain (you can follow getClassLoader().getParent()).
It also has an advantage over tomcat's WebappClassLoader by not reinventing the wheel and not depending on other objects. It re-uses code from URLClassLoader as much as possible.
(it doesn't yet honor the system class loader, but when I get that fixed I'll update the answer)
It honors the system class loader (for java.* classes, endorsed dir, etc.). It also works when security is turned on and the classloader doesn't have access to its parent (yes, this situation is weird, but possible).
public class ChildFirstURLClassLoader extends URLClassLoader {
private ClassLoader system;
public ChildFirstURLClassLoader(URL[] classpath, ClassLoader parent) {
super(classpath, parent);
system = getSystemClassLoader();
}
#Override
protected synchronized Class<?> loadClass(String name, boolean resolve)
throws ClassNotFoundException {
// First, check if the class has already been loaded
Class<?> c = findLoadedClass(name);
if (c == null) {
if (system != null) {
try {
// checking system: jvm classes, endorsed, cmd classpath, etc.
c = system.loadClass(name);
}
catch (ClassNotFoundException ignored) {
}
}
if (c == null) {
try {
// checking local
c = findClass(name);
} catch (ClassNotFoundException e) {
// checking parent
// This call to loadClass may eventually call findClass again, in case the parent doesn't find anything.
c = super.loadClass(name, resolve);
}
}
}
if (resolve) {
resolveClass(c);
}
return c;
}
#Override
public URL getResource(String name) {
URL url = null;
if (system != null) {
url = system.getResource(name);
}
if (url == null) {
url = findResource(name);
if (url == null) {
// This call to getResource may eventually call findResource again, in case the parent doesn't find anything.
url = super.getResource(name);
}
}
return url;
}
#Override
public Enumeration<URL> getResources(String name) throws IOException {
/**
* Similar to super, but local resources are enumerated before parent resources
*/
Enumeration<URL> systemUrls = null;
if (system != null) {
systemUrls = system.getResources(name);
}
Enumeration<URL> localUrls = findResources(name);
Enumeration<URL> parentUrls = null;
if (getParent() != null) {
parentUrls = getParent().getResources(name);
}
final List<URL> urls = new ArrayList<URL>();
if (systemUrls != null) {
while(systemUrls.hasMoreElements()) {
urls.add(systemUrls.nextElement());
}
}
if (localUrls != null) {
while (localUrls.hasMoreElements()) {
urls.add(localUrls.nextElement());
}
}
if (parentUrls != null) {
while (parentUrls.hasMoreElements()) {
urls.add(parentUrls.nextElement());
}
}
return new Enumeration<URL>() {
Iterator<URL> iter = urls.iterator();
public boolean hasMoreElements() {
return iter.hasNext();
}
public URL nextElement() {
return iter.next();
}
};
}
#Override
public InputStream getResourceAsStream(String name) {
URL url = getResource(name);
try {
return url != null ? url.openStream() : null;
} catch (IOException e) {
}
return null;
}
}

By reading the source code of either Jetty or Tomcat, both of which provide parent-last class loaders to implement webapp semantics.
https://github.com/apache/tomcat/blob/7.0.93/java/org/apache/catalina/loader/WebappClassLoaderBase.java
Which is to say, by overriding the findClass method in your ClassLoader class. But why reinvent the wheel when you can steal it?
Reading your various updates, I see that you ran into some classic problems with the XML SPI system.
The general problem is this: if you create a completely isolated class loader, then it's hard to use the objects it returns. If you allow sharing, you can have problems when the parent contains the wrong versions of things.
It is to deal with all this lunacy that OSGi was invented, but that's a big pill to swallow.
Even in webapps, the class loaders exempt some packages from the 'local-first' processing on the assumption that the container and the webapp have to agree on the API between them.

(see at the bottom for an update on a solution I found)
It seems that AntClassLoader has a support for parent first/last, (didn't test it yet)
http://svn.apache.org/repos/asf/ant/core/trunk/src/main/org/apache/tools/ant/AntClassLoader.java
Here is a snippet
/**
* Creates a classloader for the given project using the classpath given.
*
* #param parent The parent classloader to which unsatisfied loading
* attempts are delegated. May be <code>null</code>,
* in which case the classloader which loaded this
* class is used as the parent.
* #param project The project to which this classloader is to belong.
* Must not be <code>null</code>.
* #param classpath the classpath to use to load the classes.
* May be <code>null</code>, in which case no path
* elements are set up to start with.
* #param parentFirst If <code>true</code>, indicates that the parent
* classloader should be consulted before trying to
* load the a class through this loader.
*/
public AntClassLoader(
ClassLoader parent, Project project, Path classpath, boolean parentFirst) {
this(project, classpath);
if (parent != null) {
setParent(parent);
}
setParentFirst(parentFirst);
addJavaLibraries();
}
Update:
Found this as well, when as a last resort I started guessing class names in google (this is what ChildFirstURLClassLoader produced) - but it seems to be incorrect
Update 2:
The 1st Option (AntClassLoader) is very coupled to Ant (requires a Project context and not easy to pass a URL[] to it
The 2nd Option (from an OSGI project in google code) wasn't quite what I needed as it searched parent classloader before the system classloader (Ant class loader does it correctly by the way). The problem as I see it, think that your parent classloader includes a jar (that it shouldn't have) of a functionality that wasn't on JDK 1.4 but was added in 1.5, this has no harm as the parent last class loader (regular delegation model, e.g. URLClassLoader) will always load first the JDK's classes, but here the child first naive implementation seems to unveil the old, redundant jar in the parent class loader, shadowing the JDK / JRE own implementation.
I have yet to find a certified, fully tested, mature Parent Last / Child First correct implementation that is not coupled to a specific solution (Ant, Catalina/Tomcat)
Update 3 - I found it!
I WAS looking in the wrong place,
All I did was add META-INF/services/javax.xml.transform.TransformerFactory and restored the JDK's com.sun.org.apache.xalan.internal.xsltc.trax.TransformerFactoryImpl instead of the old Xalan's org.apache.xalan.processor.TransformerFactoryImpl
The only reason I don't "accept my own answer" yet, is that I don't know if the META-INF/services approach has the same classloader delegation as regular classes (e.g. is it parent-first / child-last or parent-last / child-first?)

URLClassLoader had this constructor public URLClassLoader(URL[], ClassLoader) that allows you to override the parent classloader of an URLClassLoader. You can just load your classloader through an URLClassLoader with an overridden parent classloader.

You can override findClass() and loadClass() to implement a child first class loader:
/**
* Always throws {#link ClassNotFoundException}. Is called if parent class loader
* did not find class.
*/
#Override
protected final Class findClass(String name)
throws ClassNotFoundException
{
throw new ClassNotFoundException();
}
#Override
protected Class loadClass(String name, boolean resolve)
throws ClassNotFoundException
{
synchronized (getClassLoadingLock(name)){
/*
* Check if we have already loaded this class.
*/
Class c = findLoadedClass(name);
if (c == null){
try {
/*
* We haven't previously loaded this class, try load it now
* from SUPER.findClass()
*/
c = super.findClass(name);
}catch (ClassNotFoundException ignore){
/*
* Child did not find class, try parent.
*/
return super.loadClass(name, resolve);
}
}
if (resolve){
resolveClass(c);
}
return c;
}
}