Hi,
The difference between null and “?” is that null binds to the first bundle that
gets injected with the pid, whereas “?” remains “?” forever. If you have two
bundles that use the same pid then null will bite you in the ass, and result in
one or other of the bundles (whichever gets injected
On Tue, Oct 18, 2016 at 3:18 PM, Paul F Fraser wrote:
> Hi Benson,
>
> Instead of null
>
> try using configurationAdmin.getConfiguration(pid, "?");
Thanks. This seems to have improved things. Now I can't explain why
the working test worked before, but you can't have
Perhaps it might help to show the DTO you've dumped here so more eyes can
look at it.
Sincerely,
- Ray
On Tue, Oct 18, 2016 at 8:46 AM, Benson Margulies
wrote:
> On Tue, Oct 18, 2016 at 8:40 AM, Timothy Ward
> wrote:
> > How are you creating the
On Tue, Oct 18, 2016 at 8:40 AM, Timothy Ward wrote:
> How are you creating the configuration?
>
> Configuration Admin uses the concept of a bundle location to restrict the
> visibility of a pid to a particular bundle. This almost always trips people
> up when they create
How are you creating the configuration?
Configuration Admin uses the concept of a bundle location to restrict the
visibility of a pid to a particular bundle. This almost always trips people up
when they create the configuration themselves (i.e. in code) as the “default”
methods bind the
Here's my puzzle.
I have a set of services that all have unsatisfied references to the
same thing - a vanilla @Reference to the interface WorkerBusService.
In each case, there is a configuration DTO with an unsatisfied
reference with name BusConfigured and target null.
WorkerBusService is