Might I suggest:

GET_ONE: "name=foo"
GET_ALL: "name=all"
services: (|(name=<name>)(name=all))

Don

On Wed, May 16, 2012 at 1:40 PM, Tim Moloney <[email protected]> wrote:
> I am having trouble figuring out how to implement the following using the
> whiteboard pattern.
>
> There are several services registered with OSGi.
> - All services provide status.
> - One service (call it GET_ONE) also registers to get the status of one
> specific service.
> - Another service (call it GET_ALL) also registers to get the status of all
> the services.
>
> So far, I have only been about to
> - have both GET_ONE and GET_ALL get all the status
> - have GET_ONE get its desired status and GET_ALL get no status
>
> I know that I must have made a mistake in the filtering but I can't figure
> out where.
>
> This is what I've tried.
>
> Scenario 1
> - All services are registered according to their function
> - GET_ALL also registered as a StatusListener with *no* service properties
> - GET_ONE also registered as a StatusListener with "name=foo" service
> property
> - When sending status, all the services use *no* filter in the query
> Results
> - GET_ALL gets all the status (good)
> - GET_ONE gets all the status (bad)
>
> Scenario 2
> - All services are registered according to their function
> - GET_ALL also registered as a StatusListener with *no* service properties
> - GET_ONE also registered as a StatusListener with "name=foo" service
> property
> - When sending status, all the services use *"name=<name>"* filter in the
> query
> Results
> - GET_ALL gets no status (bad)
> - GET_ONE gets its desired status (good)
>
> Scenario 3
> - All services are registered according to their function
> - GET_ALL also registered as a StatusListener with *"name=*"* service
> properties
> - GET_ONE also registered as a StatusListener with "name=foo" service
> property
> - When sending status, all the services use *"name=<name>"* filter in the
> query
> Results
> - GET_ALL gets no status (bad)
> - GET_ONE gets its desired status (good)
>
> Can someone point out the problem?  Thanks.
>
> Tim

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to