[snip]
Brian O'Neill wrote:
OK, per Guillaume's suggestion perhaps we start anew basing everything
on 0.90 sca.
So, what are peoples thoughts towards the design of the translation
layer?
Should we leverage Tuscany's parsing capabilities to read in the SCA
contribution?
Then, from the parsed structure generate the service-unit JBI artifacts?
Just a thought about that translation layer... If you only need the SCA
contribution and assembly models and the ability to read SCA assembly
XML, you don't have to use the whole Tuscany runtime. We've
rearchitected Tuscany SCA a few months ago to support that kind of
scenario and make it easy to reuse / embed a subset of Tuscany modules
in tools, generators, and platforms that are only interested in the SCA
metadata without dragging the whole thing.
The Tuscany modules are there:
http://svn.apache.org/repos/asf/incubator/tuscany/java/sca/modules/
For the SCA assembly, SCA contribution and policy metadata models alone,
grab these modules:
assembly
interface
contribution
policy
Support for Java and WSDL interfaces, and Java component implementations:
interface-java
interface-wsdl
implementation-java
Support for reading SCA assembly XML and handling SCA contributions:
assembly-xml
interface-java-xml (also introspects Java interfaces)
interface-wsdl-xml (also "introspects" WSDL portTypes)
implementation-java-xml
contribution-impl
These modules are self contained and should provide you with the ability
to process SCA contributions and read SCA assembly XML, without
dependencies on the Tuscany runtime, IoC container, etc. To draw an
analogy with other projects, you could compare that level of function
with packages like WSDL4J or Woden for WSDL for example: Models and the
ability to read/write them.
Hope this helps
--
Jean-Sebastien
