I agree that throwing a RuntimeException is a good solution, the
important thing is really to preserve the stack info of the original
exception, i.e. pass the original exception to the RuntimeException
constructor.
A related question that I've trying to figure out for a long time:
what is the role of the various subclasses of RuntimeException in
Rhino, and how are you supposed to use them (or not)? Here's the class
hierarchy looks like:
+-java.lang.RuntimeException
+-RhinoException
+-EcmaError
+-JavaScriptException
+-EvaluatorException
+- WrappedException
Hannes
On Nov 24, 10:53 am, Attila Szegedi <[EMAIL PROTECTED]> wrote:
> I'd say, yes. I personally dislike the checked exceptions mechanism of
> Java, and usually work around it without much sense of guilt -- if my
> implementation of some method throws a checked exception I can't
> propagate because of an
> interface constraint, I usually just throw a new
> java.lang.reflect.UndeclaredThrowableException wrapping the original
> checked exception.
>
> This works most of the time -- there might be some corner cases where
> this is undesirable (i.e. if you'd want to have specific recovery at a
> higher level, and the compiler will no longer force you to handle it
> so you need to remember it yourself), but as long as your code is
> structured so that you don't have much catch blocks but mostly finally
> blocks, and a logging catch in the top level invoker of your work unit
> (HTTP request, JMS message receipt, UI event, etc.) it's suitable, as
> the original exception gets logged and can be inspected.
>
> Attila.
>
> On 2008.11.23., at 22:01, Peter Michaux wrote:
>
> > I have a class implementing Scriptable. The "get" method needs to do
> > some disk work which may throw a checked exception. Scriptable's "get"
> > cannot throw a checked exception. What is the recommended way to pass
> > that exception through to the JavaScript? Right now I'm catching the
> > checked exception and then throwing a RuntimeException (which is not
> > checked). Is that considered a reasonable approach?
>
> > Thanks,
> > Peter
>
>
_______________________________________________
dev-tech-js-engine-rhino mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-js-engine-rhino
