Re: [cp-patches] RFC: ObjectInputStream fix for readResolve invocations

[Mark Wielaard]

Hi Wolfgang,

On Tue, 2005-11-08 at 21:55 +0100, Wolfgang Baer wrote:
> The mauve test also shows (test number 1) that readResolve is actually
> called by the ObjectInputStream implementation. Therefore I would say
> that bug 22854 (readResolve isn't called) can be closed.

Yep, you are right. I added the ChangeLog messages that fixed this to
the bug and closed it.

> 2005-11-08  Wolfgang Baer  <address@hidden>
> 
>         * java/io/ObjectInputStream.java:
>         (processResolution) Pass Error, RuntimeException and
>         ObjectStreamException through to the caller.

Yes, this makes sense.

> Index: java/io/ObjectInputStream.java
> ===================================================================
> RCS
> file: /cvsroot/classpath/classpath/java/io/ObjectInputStream.java,v
> retrieving revision 1.71
> diff -u -r1.71 ObjectInputStream.java
> --- java/io/ObjectInputStream.java      1 Nov 2005 23:32:21
> -0000       1.71
> +++ java/io/ObjectInputStream.java      8 Nov 2005 20:43:56 -0000
> @@ -1567,6 +1567,15 @@
>           }
>         catch (InvocationTargetException ignore)
>           {
> +           // SUN passes Errors, RuntimeExceptions and
> +           // ObjectStreamExceptions through to caller
> +           Throwable cause = ignore.getCause();
> +           if (cause instanceof ObjectStreamException)
> +             throw (ObjectStreamException) cause;
> +           else if (cause instanceof RuntimeException)
> +             throw (RuntimeException) cause;
> +           else if (cause instanceof Error)
> +             throw (Error) cause;

Multiple nitpicks here. The comment "Can throw Errors and
RuntimeExceptions if caused by the readResolve() user code" should not
be here, but a comment of public Object readObject() method. (The fact
that SUN also does this is not really that relevant imho.) The name
'ignore' is really a misnomer now. Maybe it isn't really confusing, but
it looks funny.

Feel free to commit this with or without these nitpicks fixed.

Thanks,

Mark

signature.asc
Description: This is a digitally signed message part