Message handlers and error handling

[Cantor, Scott]

On 4/15/14, 12:02 PM, "Tom Zeller" <tzeller at dragonacea.biz> wrote:
>
>Fine for now, every decision seems like triage.

We'll see what Brent thinks, the interface and the implementations are
largely his.

>Right ... I have been wondering if handler chains should be sub-flows
>instead. Not sure that is possible or helps us here.

I've thought about it. The advantage now, apart from just the lack of
overhead, is that they're very easy to wire together with Spring alone and
are easier to customize, but I admit it's inconsistent. Using actual
subflows would mean there'd be event transitions obviously, and the one
that actually fails would surface as the subflow result.

-- Scott