My interpretation is that a StateChangeRequired fault is not necessarily
an indication that the Consumer must explicitly clone the portlet. There
could be other reasons, ex, for security, due to which the Producer may
not allow the portlet to make state changes. So, even after cloning, the
Consumer may continue to receive the same fault.
Subbu
Richard Jacob wrote:
>
>
>
>
> Subbu Allamaraju <> wrote on 04/29/2004 04:17:57 PM:
>
> > Richard,
> >
> > My interpretation is that a Producer should not throw
> > StateChangeRequired fault when the portletStateChange flag is set to
> > cloneBeforeWrite or readWrite.
>
> Yes I was just talking about cloneBeforeWrite
>
> > Firstly, the Producer has a chance to
> > make state changes. Secondly, with the interpretation you present, the
> > consumer first makes a call with cloneBeforeWrite, gets a fault, clones,
> > and then repeats the pbia. This is more work for the Consumer and more
> > waiting time for the end user.
>
> Correct, but the spec allows that and might ease Producer implementations
> and puts a burdon on the Consumer side (what we always did).
>
> Richard
>
