hi jukka
On Thu, Apr 26, 2012 at 10:45 AM, Angela Schreiber<[email protected]> wrote:
i keep wondering what was the reason for keeping implementations
of the oak-api in different packages. specifically i don't fully
understand what was the logic behind having some classes
in the oak.core package (KernelContentSession), while others
are in the oak.kernel package (KernelPropertyState).
My original idea for the different packages was:
* .oak.kernel - type-safe wrapper (NodeState, etc.) for the
MicroKernel interface
* .oak.core - implementation of the Oak API based on .oak.kernel
Currently that distinction is a bit blurred, and I agree that it
should be fixed. However, instead of moving everything to the same
package, let's do something like this:
* Keep the MicroKernel wrappers in .oak.kernel
* Move the Oak API implementation classes to .oak.core
* Perhaps rename the classes in .oak.core from KernelSomething to
SomethingImpl to avoid confusion
so, i would like to suggest that we move the classes from the
kernel package to the core package.
if nobody objects i would go ahead and commit the move later
on today.
Does the above alternative sound good to you?
i guess that's fine with me. let me try it out
kind regards
angela
BR,
Jukka Zitting
