I'll jump in though I might be missing some context... ...

I wonder if Shaoguang is describing the fact that the component service name
must match the implementation service name when you are configuring the
component via SCDL (rather than making a statement about the composite
service name).

This issue is particularly noticable when there is only one service in a
component since in most places SCA lets you be ignorant of the service name
of a single-service-component (e.g. when making a wire target).

A "guiding exception" would be a nice addition.......

Scott



On 8/3/07, ant elder <[EMAIL PROTECTED]> wrote:
>
> According to line 1498 on page 34 of the Assembly spec[1] it doesn't sound
> like thats correct:
>
> 1498 • name (required) – the name of the service, the name MUST BE unique
> across all the
> 1499 composite services in the composite. The name of the composite
> service
> can be different
> 1500 from the name of the promoted component service.
>
> So the behaviour you're seeing sounds like a bug. Could you raise a JIRA,
> and if possible attach some code to demonstrate the problem?
>
>    ...ant
>
> [1] http://www.oasis-opencsa.org/sca-assembly
>
> On 8/3/07, shaoguang geng <[EMAIL PROTECTED]> wrote:
> >
> > Just now, I found that the <service name="[name]">, name here must equal
> > to the java interface's name, as well the service's name of WSDL, other
> > wise, we just see NullPointerException.
> >
> > I would suggest generate a guiding Exception, to tell the developer this
> > "rule" of defining a service.
> >
> >
> > ---------------------------------
> > Sick sense of humor? Visit Yahoo! TV's Comedy with an Edge to see what's
> > on, when.
>

Reply via email to