-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I have now looked at the details of adding support for a custom shutdown
method and have come up with some rouge ideas and would like some input
for all of you if I am going down the right (or wrong) path.

1) Adding a shutdown-method attribute to the builder factory.  This
would be default be null (i.e. no registration).  If specified, must be
a valid, no-param method.  This could also be a 'known' method name,
which would always be called (this could of course build up a large
number of references to services/methods).  Thoughts?
2) Lookup and store the reference to the instance and method to invoke
(in the BuilderFactoryLogic class)
3) Remove the registration for shutdown in the service models
4) Allow the builder factory to query whether or not the instance
(service model) supports shutdown.  Currently passes a boolean param to
the instance factory (as invoked from the service model).
5) Anything else?

Does this make sense?  If so, I can add an issue in Jira and attach my
updates.

Thanks,

Johan

Johan Lindquist wrote:
> Whatever the name is, I would say this is something that would polish
> off the life cycle methods support.  I'll raise an issue in Jira and
> have a peek to see what the changes required would be.
> 
> On a different note, I was wondering what everyones thought would be on
> adding an 'before', 'after' and a 'name' attribute to the EagerLoad
> contribution point (similar to the interceptor contribution)? It would
> be great to be able to control te order of initialization and this adds
> a familiar way of doing so.
> 
> Cheers,
> 
> Johan
> 
> 
> James Carman wrote:
> 
>> I probably would have used a method name like dispose() or shutdown() or
>>destroy() rather than registryDidShutdown() if you want to stress the
>>importance of removing a dependency (the "registry" part screams HiveMind).
>>But, your concern is a valid one.  We have support for init methods, but not
>>destroy/shutdown methods.
>>
>>-----Original Message-----
>>From: Johan Lindquist [mailto:[EMAIL PROTECTED] 
>>Sent: Thursday, January 05, 2006 12:09 PM
>>To: hivemind-dev@jakarta.apache.org
>>Subject: Registry shutdown
>>
>>Hi all,
>>
>>Is there a particular reason for the RegistryShutdownListener interface?
>> Could this not be implemented just as the automagic initialization call
>>of the service?  If a method on the implementation is named
>>'registryDidShutdown', add it automatically to the list of shutdown
>>listeners?  Would nicely remove another compile time dependancy.
>>
>>Cheers,
>>
>>Johan
>>
>>--
>>you too?
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
> 
> 

- --
you too?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDwkh41Tv8wj7aQ34RAq3PAJ92qsd4prChRvxqslrjWQ68nCK0+ACfViPu
muD7bTsk++eyzU94zCkHCR4=
=sHro
-----END PGP SIGNATURE-----

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to