We Keep Coding

Class.forName() is not working with a particular class

I am trying to load with java reflection a bunch of classes. Everything seems working fine (I am handling the exception if the class is not found).
However there is a particular class that is raising another exception that is not thrown by call to the Class.forname() but by an internal method and so I cannot even catch it.
Here is my code:
try {
URL url = Class.forName(qualifiednameOfTheClass);
} catch (ClassNotFoundException ex) {
// ok class not found can be handled
} catch (Exception e){
// catch every other exception just to try to get the strange exception
}
So with this code everything is working, I am using it on lots of classes and it's working (sometimes it finds it sometimes it doesn't).
However there is one case that is not working properly and I cannot understand why. If qualifiednameOfTheClass = sun.security.krb5.SCDynamicStoreConfig my code is raising an exception:
Exception in thread "mythread-1" java.lang.UnsatisfiedLinkError: no
osx in java.library.path at
java.lang.ClassLoader.loadLibrary(ClassLoader.java:1886) at
java.lang.Runtime.loadLibrary0(Runtime.java:849) at
java.lang.System.loadLibrary(System.java:1088) at
sun.security.action.LoadLibraryAction.run(LoadLibraryAction.java:67)
at
sun.security.action.LoadLibraryAction.run(LoadLibraryAction.java:47)
at java.security.AccessController.doPrivileged(Native Method) at
sun.security.krb5.SCDynamicStoreConfig.(SCDynamicStoreConfig.java:39)
at java.lang.Class.forName0(Native Method) at
java.lang.Class.forName(Class.java:190) at
com.myclass.getJARFromClassForName(JavaId.java:510)
at com.myclass.getJARUrl(Id.java:550) at
com.myclass.collectInformation(Graph.java:366)
at
com.myclass.createNode(Graph.java:166)
at com.myclass.Graph.(Graph.java:143) at
com.myclass2.run(myclass2.java:246)
at java.lang.Thread.run(Thread.java:745)
So as you can see in the error we have this strange exception that cannot be caught even with a generic catch like in my code and I cannot understand why it has been raised and what actually is this osx library (I am on linux)
EDIT: The only thing that I found is this link http://www.straub.as/java/pocketapi/index7.html but is in german and so I read it with google translate and I don't know if I got it right but is saying that the classes listed there cannot be reproduced with Class.forname() itself.
Is it true? Is there a reason why this cannot be loaded with reflection or am I doing something wrong?

"Cannot be caught" because it's an Error, not an Exception. I'd recommend reviewing the object hierarchy in the JDK for Throwable.
Try changing that to Throwable and you'll have better luck catching. I don't know why the error is happening.
This looks like a JNI class that's using native code. I don't know what you're doing, but this looks like a bad idea to me.

This is not how this class expects to be loaded and loading this internal class directly doesn't appear to work. You need to load the class using the standard encryption API so that this class can be loaded as expected, or possibly not loaded at all (it could be code which will only work on OSX)

As you can see in the link you have provided, there are some classes where this method fails.
When calling this method on these objects, Java needs first to load additional stuff, as this is platform specific stuff which is not shipped by default. In your case, it is Kerberos, a security API.
As you can see in its Documentation, it searches for some files in specific paths (java.library.path). As it can not find it there, it throws an error.
Note that the error UnsatisfiedLinkError does not refer to finding the class name for sun.security.krb5.SCDynamicStoreConfig itself. It refers to not finding the native library in the paths provided by java.library.path.
This path itself points, for example on Windows, to C:Windows/system32/.
However, you may catch this error with catch(Error e), note that an Error is not an Exception (Throwable hierarchy).
Be aware that catching an Error in general is no good idead as you can not be sure if the JVM can recover from it.

DeepLearning4J NoSuchMethodError

I'm new to neural networks and NLP. I've found this library: DeepLearning4J. I'm trying to get it to work but whenever I execute this instruction:
Collection<String> similar = vec.wordsNearest("word_to_search", 10);
If the word I'm searching is mapped into the network I get the following exception:
java.lang.IllegalArgumentException: XERBLA: Error on argument 6 (LDA) in SGEMV
at org.jblas.NativeBlas.sgemv(Native Method)
at org.nd4j.linalg.jblas.blas.JblasLevel2.sgemv(JblasLevel2.java:25)
at org.nd4j.linalg.api.blas.impl.BaseLevel2.gemv(BaseLevel2.java:53)
at org.nd4j.linalg.api.ndarray.BaseNDArray.mmuli(BaseNDArray.java:2569)
at org.nd4j.linalg.api.ndarray.BaseNDArray.mmul(BaseNDArray.java:2377)
at org.deeplearning4j.models.embeddings.wordvectors.WordVectorsImpl.wordsNearest(WordVectorsImpl.java:290)
at org.deeplearning4j.models.embeddings.wordvectors.WordVectorsImpl.wordsNearest(WordVectorsImpl.java:337)
at word2vec.Word2VecTest.main(Word2VecTest.java:74)
Exception in thread "main" java.lang.NoSuchMethodError: org.nd4j.linalg.api.ndarray.INDArray.mean(I)Lorg/nd4j/linalg/api/ndarray/INDArray;
at org.deeplearning4j.models.embeddings.wordvectors.WordVectorsImpl.wordsNearest(WordVectorsImpl.java:283)
at word2vec.Word2VecTest.main(Word2VecTest.java:89)
I know that the NoSuchMethodError may be due to libraries different versions. In this specific case, this is probably caused by nd4j. I've checked the versions lots of time and this is what I'm importing at the moment:
akka-actor_2.11-2.4-M3.jar
akka-cluster_2.11-2.4-M3.jar
akka-remote_2.11-2.4-M3.jar
akka-slf4j_2.11-2.4-M3.jar
byte-buddy-0.6.15.jar
config-1.3.0.jar
deeplearning4j-core-0.0.3.3.4.alpha2.jar
deeplearning4j-nlp-0.0.3.3.4.alpha2.jar
deeplearning4j-scaleout-akka-0.0.3.3.4.alpha2.jar
deeplearning4j-ui-0.0.3.3.4.alpha2.jar
javassist-3.12.1.GA.jar
jblas-1.2.4.jar
jcublas-6.5.jar
lucene-analyzers-common-4.10.3.jar
lucene-core-4.10.3.jar
nd4j-api-0.4-rc3.4.jar
nd4j-bytebuddy-0.4-rc3.4.jar
nd4j-jblas-0.4-rc3.4.jar
nd4j-jcublas-common-0.4-rc3.4.jar
netty-3.10.4.Final.jar
protobuf-java-2.6.1.jar
reflections-0.9.10.jar
scala-library-2.12.0-M2.jar
selenium-server-standalone-2.47.1.jar
Can someone explain to me the problem?

The error is telling you that DeepLearning4J tried to call the method INDArray INDArray.mean(int value) but this method was not found.
Looking at nd4j 0.4-rc3.4 source code, you can see that the mean method actually takes a vararg int... as input. Since this is not int, the error is thrown.
This change was made by this commit when nd4j bumped version from 0.0.3.5.5.5 to 0.4-rc0.
As a result, you need to downgrade nd4j to version 0.0.3.5.5.5. With this downgrade, you will not have any more incompatibility since this is the actual version that DeepLearning4J is depending on. You can see that in the Maven dependencies of deeplearning4j-core-0.0.3.3.4.alpha2.

How can I tell why I'm getting an IOException at line 113 in com.sun.jdmk.comm.HttpSendSocket

My JDMK based application is getting intermittent IOExceptions at line 313 in com.sun.jdmk.comm.HttpSendSocket and I can't figure out why. All I know from the Javadoc about this is if an I/O error occurs while creating the input stream you'll get an IOException, but I don't know what kind of I/O error occurred or why one did. The code actually worked both before and after this error transpired.
Any tips on how to debug this intermittent problem would be appreciated.
I don't want to paste the source code here for HttpSendSocket since it belongs not to me, but I know it's doing an HttpURLConnection conn.getInputStream() when the IOException exception is thrown.
I thought about trying to create my own version of HttpSendSocket, and adding diagnostics in it, but couldn't figure out how since it's a package protected class.
Stack trace below:
com.sun.jdmk.comm.CommunicationException: java.io.IOException: HTTP request failed
at com.sun.jdmk.comm.HttpSendSocket.readNotify(HttpSendSocket.java:313)
at com.sun.jdmk.comm.HttpSendInputStream.read(HttpSendInputStream.java:95)
at java.io.FilterInputStream.read(FilterInputStream.java:94)
at java.io.PushbackInputStream.read(PushbackInputStream.java:150)
at com.sun.jdmk.comm.GenericHttpConnectorClient.sendHttp(GenericHttpConnectorClient.java:486)
at com.sun.jdmk.comm.GenericHttpConnectorClient.invokeRemoteOperation(GenericHttpConnectorClient.java:2234)
at com.sun.jdmk.comm.GenericHttpConnectorClient.invoke(GenericHttpConnectorClient.java:1366)
As I said, any helpful suggestions would be appreciated.

The communication Exception was caused by using ArrayList method subList. ArrayList is serializable but subList data IS NOT serializable, and therefore you cannot retrieve the data over an HttpConnector. The solution was to change:
List<UserProcessInfo> values = new ArrayList<UserProcessInfo>();
...
values.size() <= 1000 ? values : values.subList(0,1000);
to:
List<UserProcessInfo> values = new ArrayList<UserProcessInfo>();
...
return values.size() <= 1000 ? values : new ArrayList<UserProcessInfo>(values.subList(0,1000));

Attach the JDMK source to your IDE, then set a breakpoint in the HttpSendSocket and run it with debugging enabled. At least in IntelliJ you can attach the library source by trying to open the class from the stacktrace, then choosing to link source. Don't know how the process is for other IDE's but I would expect it to be possible.

NetBeans and stop/break on all Exceptions?

I'm using NetBeans as my IDE for developing Android.
However, when I get an exception, Netbeans does not break on the exception as I expect it to.
I have checked the box "Stop on uncaught exceptions" that can be found in Options --> Java Debugger --> Stop on uncaught exceptions but that doesn't help.
Furthermore, where can I see the actual exception message? I don't see anything. I have no clue where and when an exception occurs, just that it does occur.
I've read some on the netbeans.org site about a bug in 6.9.1 that was fixed, but it doesn't seem to be fixed in 7.0 that I have.
The debugging window doesn't say anything useful at all, gives some form of stack trace that is pointless as it doesn't specify any of my own code.
I switched from Eclipse because that IDE sucks, NetBeans is much leaner, but the debugging needs to be fixed to be useful.

I have this problem when I use netbeans too. To get the exception message, I use try-and-catch. In the catch() statement, call a function that contains some code and put a breakpoint on it. Then when the exception is given your breakpoint in the catch statement will be called and you can read the information (error message, stack, etc.) from the Exception obect.
try
{
// Doing something here
}
catch (Exception e1)
{
// This is called when an exception occurs
doSomething(); /// Put breakpoint here
}
UPDATE:
I am not sure if this will work for android but you could try entering a "New Breakpoint". When entering a new breakpoint to break on an exception, you need to know the
exact package/class exception name. For example, if you want to catch a
NullPointerException, then you go to Debug >> New Breakpoint to create a New Breakpoint and enter
java.lang.NullPointerException into the Exception Class Name field in the
Breakpoint Properties dialog. Choose whether to break on caught, uncaught or both, exceptions and it will hit a breakpoint if that type of exception ever occurs in the class/project.

I had such behavior where I had my source file that generated the exception NOT in "default package". When I moved everything to "default package" it started working fine and stop on exception automatically.

This answer actually also answers this question:
https://stackoverflow.com/a/2671390/415551
To quote user Finbarr:
Go to debug > New Breakpoint (alternatively CTRL+SHIFT+F8). Change the
breakpoint type to Exception in the top right hand drop down menu.
Type java.lang.NullPointerException in the Exception class field.
Choose whether to break on caught, uncaught or both.
Debug your code and watch the glorious auto breakpoint when the
Exception is thrown.
Simply change java.lang.NullPointerException to java.lang.Exception to break on any error.

IllegalStateException: Can't overwrite cause in Guava map.put

I create a map using
new MapMaker().softValues().maximumSize(cacheSize).makeMap();
This seemed to work fine, however, I sometimes get the following exception after accessing right after deploying on the server and putting new elements into the map:
java.lang.IllegalStateException: Can't overwrite cause
at java.lang.Throwable.initCause(Throwable.java:320)
at org.jboss.mx.loading.RepositoryClassLoader.findClass(RepositoryClassLoader.java:624)
at java.lang.ClassLoader.loadClass(ClassLoader.java:307)
at org.jboss.mx.loading.RepositoryClassLoader.loadClassImpl(RepositoryClassLoader.java:474)
at org.jboss.mx.loading.RepositoryClassLoader.loadClass(RepositoryClassLoader.java:415)
at java.lang.ClassLoader.loadClass(ClassLoader.java:248)
at com.google.common.collect.CustomConcurrentHashMap$2.iterator(CustomConcurrentHashMap.java:828)
at java.util.AbstractCollection.remove(AbstractCollection.java:241)
at com.google.common.collect.CustomConcurrentHashMap$Segment.removeFromChain(CustomConcurrentHashMap.java:2599)
at com.google.common.collect.CustomConcurrentHashMap$Segment.processPendingCleanup(CustomConcurrentHashMap.java:2772)
at com.google.common.collect.CustomConcurrentHashMap$Segment.runLockedCleanup(CustomConcurrentHashMap.java:2860)
at com.google.common.collect.CustomConcurrentHashMap$Segment.preWriteCleanup(CustomConcurrentHashMap.java:2806)
at com.google.common.collect.CustomConcurrentHashMap$Segment.put(CustomConcurrentHashMap.java:2374)
at com.google.common.collect.CustomConcurrentHashMap.put(CustomConcurrentHashMap.java:3346)
at my.app.cache.CacheImplGoogleGuava.put(CacheImplGoogleGuava.java:36)
...
What could be the reason?
--- Updated:
JBoss version is 5.
Setting a breakpoint in Throwable.initCause revealed ClassNotFoundException with message:
Invalid use of destroyed classloader for com.google.common.collect.Iterators, UCL destroyed at:
and with Stacktrace
ClassNotFoundException(Throwable).initCause(Throwable):320
UnifiedClassLoader3(RepositoryClassLoader).findClass(String):628
...
UnifiedClassLoader3(ClassLoader).loadClass(String):248
CustomConcurrentHashMap$2.iterator():828
CustomConcurrentHashMap$2(AbstractCollection<E>).remove(Object):241
CustomConcurrentHashMap$Segment.enqueueCleanup(...):2738
CustomConcurrentHashMap$Segment.unsetValue(...):2662
CustomConcurrentHashMap<K, V>.reclaimValue(...)
CustomConcurrentHashMap$SoftValueReference<K, V>.finalizeReferent():1637
...
Method.invoke:574
Finalizer.claenUp:154
Finalizer.run:127
From the stacktrace, it seems as if an object in the map was finalized where in finalizeReferent the class com.google.common.collect.Iterators cannot be loaded.

Your problem seems to be unrelated to MapMaker or Guava in general.
You only see an exception that happens while handling another exception (unfortunately).
Throwable.initCause() throws an exception when it's called while there was already a cause specified for the current Throwable (either via the method or the constructor`.
The RepositoryClassLoader.findClass() method seems to be handling some exception that it expects not to have a cause, but in fact it already has a cause set, which triggers this exception.
Unfortunately the exception you see here hides the actual exception (which would probably be much more important for solving the problem).
Try to put a breakpoit at Throwable.initCause() at line 320 or at RepositoryClassLoader.findClass() (at line 624) and reproduce the problem to (hopefully) see the "real" exception in the local variables view of your IDE.

This seems to be a known JBoss bug that might have already been resolved in new versions of the platform. Here is at least one similar bug:
JBREM-552: cannot init cause of ClassCastException
You might try upgrading your version of JBoss (though this particular bug implies the fix was in 2006) or filing another bug if your version is up to date. It could be the fix was sloppy and only fixed the one case that the author of that bug reported, and not similar cases.