[Registry-dev] Resource lifecycle handling

[Deepal Jayasinghe]

>
> Could you explain a little more here?  I don't see how it makes 
> development any easier to copy every time there's a transition - a 
> transition, as I see it, just calls some code which can do whatever it 
> wants.  
Yes we should provide that but what I am telling here is our default 
life cycle behavior should be something similar to SVN branch. If you 
take a  real system as an example (say SOA) with some WSDL then we do 
the transition there are high probability that we change the port 
address , so in such cases coping the content make sense.
>
> I'd rather give them the choice, like we give them the choice where to 
> put their WSDLs.  Any given way we might constrain users into doing 
> this would be inappropriate for someone out there, IMHO.
Well , I am also saying the same thing ,giving chance to invoke custom 
code mean we are giving full control.
>
> Sure - for instance, every resource on the system might be in a 
> lifecycle involving editorial review to watch for spelling and 
> profanity which has "needs-review" and "ok" states.  In addition, 
> different departments might have their own development lifecycles of 
> the built/tested/deployed variety.
Well as I can understand you are talking lifecycle + lifecycle 
conditions.  So in any given lifecycle phase there may be many stages 
(or conditions), but at any given time a resource can not be in two 
lifecycle phases. As an example for QA state to be satisfied , there 
might have a number of scenario checking (different JDKs , different OS 
and etc... ) , so once all done it will move to some other state (say 
production ). Then the resource is in production state.

-Deepal