No StackTrace in java.lang.NullPointerException

[Mr Smith]

Hello,

in my app, i catch a java.lang.NullPointerException but the stackTrace is empty:

"ex"= NullPointerException (id=44)
cause= NullPointerException (id=44)
detailMessage= null
stackTrace= null

It's hard to debug whith this. Is there any option to pass to the JVM to force setting the stack trace? Or Does it come from a bad throw ? (i haven't wrote all the source code i'm working on), or from a bad catch?

Thx for your help

 

[Mr Smith]

Hello,

in my app, i catch a java.lang.NullPointerException but the stackTrace is empty:

"ex"= NullPointerException (id=44)
cause= NullPointerException (id=44)
detailMessage= null
stackTrace= null

It's hard to debug whith this. Is there any option to pass to the JVM to force setting the stack trace? Or Does it come from a bad throw ? (i haven't wrote all the source code i'm working on), or from a bad catch?

some Infos:
java.runtime.name = Java(TM) 2 Runtime Environment, Standard Edition
java.vm.version = 1.4.2_06-b03
java.vm.vendor = Sun Microsystems Inc.
java.vm.name = Java HotSpot(TM) Client VM
java.runtime.version = 1.4.2_06-b03
java.awt.graphicsenv = sun.awt.X11GraphicsEnvironment
os.arch = i386
os.name = Linux
java.class.version = 48.0
os.version = 2.6.3-7mdk-i686-up-4GB
file.encoding = ISO-8859-15
java.specification.version = 1.4
java.vm.specification.version = 1.0
sun.arch.data.model = 32
java.vm.info = mixed mode
java.version = 1.4.2_06
sun.cpu.endian = little


Thx for your help

 

[Thomas Schodt]

Hello,

in my app, i catch a java.lang.NullPointerException but the stackTrace is empty:

"ex"= NullPointerException (id=44)
cause= NullPointerException (id=44)
detailMessage= null
stackTrace= null

It's hard to debug whith this. Is there any option to pass
to the JVM to force setting the stack trace? Or Does it come
from a bad throw ? (i haven't wrote all the source code i'm
working on), or from a bad catch?

On the assumption that you are using some library
that the owner is trying to keep you from peeking into
I tried this:

class Secret {
private static void m1() {
m2();
}
private static void m2() {
NullPointerException ex = new NullPointerException();
ex.setStackTrace(new StackTraceElement[0]);
throw ex;
}
public static void main(String[] arg) {
try {
m1();
} catch (NullPointerException ex) {
System.out.println(ex.getCause());
System.out.println(ex.getMessage());
System.out.println(ex.getStackTrace());
}
}
}

But the stack trace is an empty StackTraceElement vector (not null).

I guess someone could be doing this with JNI?


One option is to wrap any external classes / methods and throw a new
exception (there are exceptions that will allow you to wrap the original
exception) so at least you know where you are in your code when the
exception occurs.


All this assuming you have not found a bug in the VM (unlikely).

 

[Mr Smith]

Hello,

in my app, i catch a java.lang.NullPointerException but the stackTrace is empty:

"ex"= NullPointerException (id=44)
cause= NullPointerException (id=44)
detailMessage= null
stackTrace= null

It's hard to debug whith this. Is there any option to pass to the JVM to force setting the stack trace? Or Does it come from a bad throw ? (i haven't wrote all the source code i'm working on), or from a bad catch?

Some it was a bug in the JVM:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=5098422

I've downloaded the last version of 1.4.2:

java.runtime.name = Java(TM) 2 Runtime Environment, Standard Edition
java.vm.version = 1.4.2_07-b05
java.vm.vendor = Sun Microsystems Inc.
java.vm.name = Java HotSpot(TM) Client VM
java.vm.specification.name = Java Virtual Machine Specification
java.runtime.version = 1.4.2_07-b05
java.awt.graphicsenv = sun.awt.X11GraphicsEnvironment
os.arch = i386
os.name = Linux
java.class.version = 48.0
os.version = 2.6.3-7mdk-i686-up-4GB
file.encoding = ISO-8859-15
java.specification.version = 1.4
java.vm.specification.version = 1.0
sun.arch.data.model = 32
java.specification.vendor = Sun Microsystems Inc.
java.vm.info = mixed mode
java.version = 1.4.2_07
java.vendor = Sun Microsystems Inc.

i've set the cmd line argument:

-XX:-OmitStackTraceInFastThrow


but it still doesn't work...

I think i'm gonna report a bug.

 

[Mr Smith]

[...]

On the assumption that you are using some library
that the owner is trying to keep you from peeking into
nope


I guess someone could be doing this with JNI?
nope


All this assuming you have not found a bug in the VM (unlikely).

well...

 

[Roland]

Hello,

in my app, i catch a java.lang.NullPointerException but the stackTrace is empty:

"ex"= NullPointerException (id=44)
cause= NullPointerException (id=44)
detailMessage= null
stackTrace= null

It's hard to debug whith this. Is there any option to pass to the JVM to force setting the stack trace? Or Does it come from a bad throw ? (i haven't wrote all the source code i'm working on), or from a bad catch?

Thx for your help

AFAIK there no option to force the JVM to generate a stacktrace.
Normally an exception does contain a stacktrace. However if your code
(or the code you're using) uses JNI, then some native library could have
triggered the NPE without properly filling a stacktrace or even a detail
message. (I doubt the native libraries of the Java JRE are doing this).

Another cause might be an OutOfMemoryError, preventing the JVM to
generatate a proper stack trace. The times that I was able to catch an
OutOfMemoryError, it never contained a stacktrace.
--
Regards,

Roland de Ruiter
___ ___
/__/ w_/ /__/
/ \ /_/ / \

 

[Andrea Desole]

If you are using Eclipse, you can go to Run->Add Java Exception
Breakpoint and set a breakpoint for NullPointerException. Then you can
stop the execution when the exception is thrown.
If you are not using Eclipse, then you should use it
I'm curious to see if this is a JVM bug.

 

[Mr Smith]

If you are using Eclipse, you can go to Run->Add Java Exception
Breakpoint and set a breakpoint for NullPointerException. Then you can
stop the execution when the exception is thrown.
If you are not using Eclipse, then you should use it
I'm curious to see if this is a JVM bug.

It's solved, but i don't know where it came from.

No we print the stack trace somewhere else, in a different way (no toString()) and it works. I can't explain why.

The good thing is that no i understand how the JV M handle stack traces of exceptions:
- throw => save the stacktrace in the JVM (can't be reach)
- print => build an array in java memory from the JVM memory and print it

Thx for your help