[jboss-dev] VFS Abstraction Work in EJB 3.0

[John Bailey]

No problem.  I can take care of both the VFS2 and VFS3 if desired.  They should only vary by the names of the methods called for delegation.  

Another thing to think about is the use of the VirtualFileAdapter org.jboss.injection.ServiceRefInjector.  There is a VirtualFileAdapter being created taking a VirtualFile.  This will not work once the SPI is in place as the ejb3-core project will no longer have any VFS classes available.  There reason I bring this up is this class is provided by the jboss-metadata projects, and those projects will likely either support VFS2 or VFS3, so the creation of the VirtualFileAdapter will need to be done in a way that can support either.  There is an additional VirtualFileAdapter constructor that supports a URL and path pair as well, but since the semantics of what the 'path' means in VFS have changed between the versions, this will likely be problematic.  An alternative is to avoid the VirtualFileAdapter and have the SPI VirtualFile interface extend the org.jboss.wsf.spi.deployment.UnifiedVirtualFile interface, which should remove the need to have VFS version based VirtualFileAdapter creation.  It would only require two additional methods (toURL and findChild) to support that interface, which both exist in each of the VFS versions (with a slightly different name).

John


On Feb 8, 2010, at 10:30 AM, Andrew Lee Rubinger wrote:

> Looks like Carlo got ambitious and went for the SPI. :)
> 
> I'm looking at a custom Enforcer rule as discussed in 
> http://community.jboss.org/thread/147784 today.
> 
> IMO it makes the most sense for us to stick with the EJB3 bits 
> (separating the SPI and aligning the pluggable impls), leaving you to 
> create our VFS3 impl, if you wouldn't mind.
> 
> S,
> ALR
> 
> On 02/08/2010 10:35 AM, John Bailey wrote:
>> I just wanted to make sure we are all on the same page with regard to the development of the VFS abstraction in EJB 3.0.  After our discussions last week, I was under the impression I would be doing the development of the abstraction.  This morning there was an initial commit from Carlo.  Am I still correct in assuming I will be working on this?
>> 
>> --
>>   John Bailey
>>   JBoss by Red Hat
>> --
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> jboss-development mailing list
>> jboss-development at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jboss-development
> 
> -- 
> Andrew Lee Rubinger
> Sr. Software Engineer
> JBoss by Red Hat
> http://exitcondition.alrubinger.com
> http://twitter.com/ALRubinger
> _______________________________________________
> jboss-development mailing list
> jboss-development at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jboss-development

--
  John Bailey
  JBoss by Red Hat
--