We allowed for non-namespaced parameters but I don't believe we ever actually used it.
Also in doing this we got the concept of a decorator object completely the wrong way up so the code for "do we namespace or don't we" is super confusing and this use of decorator object doesn't provide the flexibility in design that it should have done.
Removing the extra layer will leave us with less code and one fewer unused option to understand (so they're namespaced, full stop), and remove the borked decorator layer. I'm not even certain the sysparams editor code can handle the non-namespaced option (though I could be wrong).
We need to audit projects using the upgraded FB11 to make sure none of them are using the non-namespaced option first, however. I think that's a case of looking at their hive configs and checking what brain is registered to provide the 'sysparams::manager' service.
We allowed for non-namespaced parameters but I don't believe we ever actually used it. Also in doing this we got the concept of a decorator object completely the wrong way up so the code for "do we namespace or don't we" is super confusing and this use of decorator object doesn't provide the flexibility in design that it should have done.
Removing the extra layer will leave us with less code and one fewer unused option to understand (so they're namespaced, full stop), and remove the borked decorator layer. I'm not even certain the sysparams editor code can handle the non-namespaced option (though I could be wrong).
We need to audit projects using the upgraded FB11 to make sure none of them are using the non-namespaced option first, however. I think that's a case of looking at their hive configs and checking what brain is registered to provide the 'sysparams::manager' service.