[wildfly-dev] JCA service/model names a.k.a WFLY-1656

Bartosz Baranowski bbaranow at redhat.com
Wed Aug 7 05:22:32 EDT 2013



----- Original Message -----
> From: "Jesper Pedersen" <jesper.pedersen at jboss.org>
> To: wildfly-dev at lists.jboss.org
> Sent: Wednesday, August 7, 2013 11:04:40 AM
> Subject: Re: [wildfly-dev] JCA service/model names a.k.a WFLY-1656
> 
> On 08/06/2013 11:55 AM, Jason Greene wrote:
> > I remember glancing at this one with Carlo when we were in Newcastle. The
> > problem is that when there are duplicate names an id is generated. However
> > the problem is that that ID is generated using a static map. It looks like
> > everything is fine in standalone because there is only one subsystem to

Not quite. If you add activation of RA in config it will change IDs of subsequent services/model names. Also if you do this from CLI/GUI level
you cant determine which ID will be assigned ( ie. if one of activation in chain has been removed )

> > parse. However domain mode has multiple profiles, so you have multiple JCA
> > subsystems that get parsed, and therefore you get nondeterministic
> > generation. Also I believe I spotted a problem with the serialization
> > code.

True, its even worse if you consider what I've written above.

> >
> > The general things that need to happen are
> >
> > 1. Get rid of static state, and instead store the state relative to the
> > subsystem block
> > 2. Double check the marshaling
> > 3. Console needs to be updated to take an "id" on creation (HAL issue
> > needed for that one)

Console fix is there, Heiko is just waiting for this issue to be fixed. AS7 has the ID parameter and from discussion Ive had, its should be perfectly fine to enforce use of it.

> >
> 
> Ok. However, there is a customer request precisely on this attribute
> that needs to be taken into account.
> 
> Best regards,
>   Jesper
> 
> 
> _______________________________________________
> wildfly-dev mailing list
> wildfly-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/wildfly-dev
> 


More information about the wildfly-dev mailing list