Re: java.lang.Class.newinstance(): exceptions; desired behavior?

[Steven Augart]

Archie Cobbs wrote:
> Steven Augart wrote:
>
> > I think our implementation of java.lang.Class's newinstance() is
> > incorrect.  Specifically, Class.newinstance() uses reflection to
> > invoke a class's zero-argument constructor and then, if it catches an
> > InvocationTargetException, unwraps that exception, throwing whatever
> > exception was originally thrown by the constructor.
> >
> > The exact code we currently have is:
> >
> >    catch (InvocationTargetException e)
> >      {
> >         VMClass.throwException(e.getTargetException());
> >         throw (InternalError) new InternalError
> >           ("VMClass.throwException returned").initCause(e);
> >      }
> >
> > This has the effect that Class.newinstance() can throw any exception
> > at all, without any warning to the caller.  
>
>
> This is a well-known bug in the spec.. see
>
>     http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4233093
>
> Sun has yet to do anything about it.
>
> IMHO in this case compatibility is best, however repulsive.

I have read the text of this, Archie, and I have to agree that we need 
to maintain compatibility (however repulsive) with the JDK.  I'll 
prepare a separate doc fix patch to java.lang.Class so that this issue 
doesn't confuse others.

It is now the documented behaviour as of JDK 1.5.0rc1.  (And it is 
*really annoying* to implement in a VM written in Java, like Jikes RVM).

Thanks for the quick reply, Archie.

--Steve Augart
--
Steven Augart

Jikes RVM, a free, open source, Virtual Machine:
http://oss.software.ibm.com/jikesrvm