On Jun 4, 2012, at 10:45 AM, Felix Meschberger wrote:

> Hi,
> 
> Smaller things first ...

good idea :-)

> Will you create issues for them ?
> 
> Am 03.06.2012 um 20:21 schrieb David Jencks:
> 
>> In other DS issues:
>> 
>> I'd like to separate the felix specific non-spec ComponentFactory behavior 
>> (where it acts like a ManagedServiceFactory) into a separate class so I can 
>> understand the spec compliant behavior more easily.
> 
> Basically this non-compliant behaviour is AFAICT only apparent in two 
> methods. Not sure whether factoring this out is worth it.

See https://issues.apache.org/jira/browse/FELIX-3536 Unfortunately the patch 
also contains detemining the satisfied state from the component manager class 
rather than a switch statement.
> 
>> 
>> I think that the behavior of ServiceFactory components that are configured 
>> by config admin is wrong.  AFAICT for service factory components we will 
>> never call a modify method but always destroy and recreate the instances.
> 
> Right.
> 
> Unlike other component managers updating configuration, ServiceFactory 
> components may internally have more than one actual instance to update. The 
> getInstance() method is not appropriate for this situation.
> 
> We'd probably need new methods for this situation (or some more abstract 
> mechanism) in the ImmediateComponentManager.modify():
> 
> (1) instead of the getInstance()==null check a method return true or false 
> indicating whether modify makes sense. The ServiceFactoryComponentManager 
> implementation returns true if at least one instance has been created
> 
> (2) instead of directly calling the modify method (step 4) a method is 
> implemented which is called by the ServiceFactoryComponentManager.

https://issues.apache.org/jira/browse/FELIX-3534  I'll look at how to fix this 
once I have a better idea about the concurrency problems.... or sooner.
> 
>> 
>> I also have a few cleanup changes such as better javadoc and removing stray 
>> references to no-longer-present *ing transient states that I don't think 
>> will be controversial so I plan to just commit them.
> 
> +1

https://issues.apache.org/jira/browse/FELIX-3535  I've applied the ones I have 
so far.... I'll probably have more soon.

thanks!
david jencks

> 
> Regards
> Felix

Reply via email to