Re: Cornerstone / AltRMI

Peter Royal wrote:

> On Monday, December 23, 2002, at 01:22  PM, Stephen McConnell wrote:
>
>> Actually - the easiest thing is if I commit the stuff that has 
>> already been tested and validated.  This is basically a bunch of 
>> .xtype and .xprofile files that coexists along side the component 
>> classes. The packaging is nothing more than including the xfile in 
>> the jar file. I'm drop these in - so you can take a look at what I'm 
>> talking about.
>
>
> Is it possible to generate these files from either the phoenix 
> descriptor or via a separate generation process from the original source?
>
> Or what are the drawbacks of only having a single descriptor (the 
> phoenix one) ?
> -pete
>

The source based generation of descriptors based on @phoneix tags would 
require a bunch of translators that are not in place today.  There not 
in place because basically I'm not 100% happy with either info or meta 
structures.  Problem is that there is a mix of implementation and 
functional logic in the defintions that needs to be resolved.

Before that there are issues to be resolved:

   * ability to declare depedencies independenly of the servicing 
lifecycle stage
      (e.g. a context entry that is established as a depedency)
   * simplification of the declaration fomats
   * resolution of standard naming for common context entries

For the moment the easiest thing is to include multiple descriptors - we 
are talking about 7 xfiles + 2 lines in the build file (at least that's 
what was put in place and validated against the james deployment which 
is cornerstone intensive).  The ideal senario is that we resolve a 
common descriptor format. For that - I like Leo's one step at a time 
approach to resolution!

Cheers, Steve.

>
> -- 
> To unsubscribe, e-mail:   
> <mailto:[email protected]>
> For additional commands, e-mail: 
> <mailto:[email protected]>
>
>
>

-- 

Stephen J. McConnell
mailto:[email protected]
http://www.osm.net