On Mon, Jan 26, 2009 at 1:31 PM, ant elder <[email protected]> wrote:

>
>
> On Mon, Jan 26, 2009 at 12:59 PM, Simon Laws <[email protected]>wrote:
>
>>
>>
>> On Mon, Jan 26, 2009 at 12:20 PM, ant elder <[email protected]> wrote:
>>
>>> The current Tuscany Webapp support that we have in 1.x is described on
>>> the wiki page at [1], I'd like to start bringing this up in the 2.x code
>>> base so we need to decide what we want to support. Right now i'm focusing on
>>> similar support as we have in 1.x where the webapp is a container for a
>>> Tuscany runtime and not so much on the webapp as an SCA contribution which
>>> requires deep integration into the servlet container (which we're working on
>>> separately with the Tuscany-Geronimo plugin work).
>>>
>>> So in 1.x we support three ways of defining the composites included in
>>> the webapp:
>>>
>>> - using a META-INF/sca-contributions.xml
>>> - by including deployable composite files within a folder named
>>> META-INF/sca-deployables
>>> - by using an WEB-INF/web.composite application composite
>>>
>>> How about dropping those first two and only supporting web.composite in
>>> 2.x?
>>>
>>>    ...ant
>>>
>>> [1]
>>> http://cwiki.apache.org/confluence/display/TUSCANY/SCA+Java+Web+Applications
>>>
>>
>> I'd be happy to let 2 go, i.e META-INF/sca-deployables, as this is a
>> Tuscany extension. Why would we get rid of support for
>> META-INF/sca-contributions.xml?
>>
>> Simon
>>
>
> I'm not sure about META-INF/sca-contributions.xml, was throwing it out
> there to see what people thought :)
>
> A META-INF/sca-contributions.xml file is really saying that the webapp .war
> file is an SCA contribution that can be contributed to an SCA runtime, but
> the way we run the webapps with Tuscany embedded thats not quite how we
> handle it. So one thing we could do is just not support it and just support
> local assembly within the webapp with the web.composite application
> composite (section 5.1.1 and 5.1.2 of the WD05 of the JEE spec). That
> simpler than supporting them both but there's probably pro's and con's.
>
>    ...ant
>
>
>
>
>
>
>
> I think we are treating webapp.war as a contribution. Hence
sca-contributions.xml should at least be optional. I could live without it
but it may be more work to turn it off than use it to identfy which
composite(s) should run.

I added a host-webapp-calculator sample that might help to push support for
this. Fell free to use/abuse as you see fit.

Simon

Reply via email to