On 5/23/07, Simon Laws <[EMAIL PROTECTED]> wrote:
Hi Just started looking at this in the context of bringing the aggregator sample to the Java runtime. I like the idea of having a dynamic interface as an option as it reduces the amount of configuration. The term dynamic term confused me to start with - I initially thought this was about constructing scripts with dynamic invocation style operations rather than the about the relationship between the SCA model and the defined script interface.
Very open to alternative names, an 'any' interface was also suggested previously is that better? or something else? Not really deeply into this yet but a question to start with. Is there any
way with the script container you are using of introspecting the methods that scripts provide?
Not yet no, currently neither JSR-223 or BSF really provide any introspection capabilities. I'd like to get this capability added to BSF. We could do this on a language by language basis doing language specific things in Tuscany and eventually moving that to BSF, so if you have a specific language you'd like to get working with introspection go for it, and I'd be happy to help. One problem with introspection and dynamic languages is that as things can be handled dynamically at runtime there's not necessarily going to be anything there to introspect, so ideally most things should work without requiring introspection. The latest Groovy beta release now supports annotations, I'd really like to get that working with Tuscany so it supports the SCA annotations spec, we'd need introspection to support that. Probably be easiest to do this in a separate <implementation.groovy> module (at least to start with), be a great thing to look at if anyone is interested in having a go. ...ant
