You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@avalon.apache.org by Peter Donald <pe...@apache.org> on 2002/07/06 06:27:44 UTC

Generate MetaInfo or not?

Hi,

I am at the stage where it is time to decorate some of the components with 
metadata tags. ie Go through some of the components (specifically the 
xml/monitor ones) and add tags in like;

@avalon.component lifestyle="ThreadSafe"
@avalon.service interface="FooInterface"
@avalon.dependency interface="FooInterface"
etc

This will be used to generate the coresponding metainfo files. I guess my 
question is

(1) Should I generate the descriptors and check them into CVS. Advantages 
of this approach is that it does not alter how they are currently used. 
Disadvantage is that it could be a bit painful to maintain over time.
(2) Should I put in ant tasks to generate the metainfo as part of build 
process. The advantage of this is that it is self-maintaining. The 
disadvantage is that it makes a whole bunch of these projects dependent on 
containerkit (and all its dependencies) when they otherwise would not be 
dependent.

Opinions?


Cheers,

Peter Donald
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"Faced with the choice between changing one's mind,
and proving that there is no need to do so - almost
everyone gets busy on the proof."
              - John Kenneth Galbraith
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>


Re: Generate MetaInfo or not?

Posted by Leo Simons <le...@apache.org>.
On Sat, 2002-07-06 at 06:27, Peter Donald wrote:
> Hi,
> 
> I am at the stage where it is time to decorate some of the components with 
> metadata tags. ie Go through some of the components (specifically the 
> xml/monitor ones) and add tags in like;
> 
> @avalon.component lifestyle="ThreadSafe"
> @avalon.service interface="FooInterface"
> @avalon.dependency interface="FooInterface"
> etc
> 
> This will be used to generate the coresponding metainfo files. I guess my 
> question is
> 
> (1) Should I generate the descriptors and check them into CVS. Advantages 
> of this approach is that it does not alter how they are currently used. 
> Disadvantage is that it could be a bit painful to maintain over time.

I'd say no. rule: "stuff that is/can be generated does not go into CVS";
don't think we should deviate here.

> (2) Should I put in ant tasks to generate the metainfo as part of build 
> process. The advantage of this is that it is self-maintaining. The 
> disadvantage is that it makes a whole bunch of these projects dependent on 
> containerkit (and all its dependencies) when they otherwise would not be 
> dependent.

how about adding a check whether containerkit is available, and
generating metainfo only if it is?

> Opinions?

once the metainfo stuff is stable, I have no problems with depending on
it everywhere. We have a compile-time dependency; generated jars and
binary distributions can ship without containing or needing
containerkit.

- Leo



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>