the dependence was introduced because it is needed to inform the system about certain system changes (concretely, method moved and method repackaged). of course, we can refactor it (in 3.0)
On Mar 5, 2013, at 8:18 PM, Stéphane Ducasse <[email protected]> wrote: > > On Mar 5, 2013, at 9:53 AM, Pavel Krivanek <[email protected]> wrote: > >> Hi, >> >> the problem is that you have added the new dependency on RPackage to >> the Kernel, namely to >> ClassDescription>>addAndClassifySelector:withMethod:inProtocol:notifying:. > > I think that RPackage should be in the core and we should remove categories. > > And UsersManager should not be in the image. > >> >> I solved it with a next little hack in the loading script but we >> should rethink the role in RPackage in the kernel. And I think we >> should make our beta policy stricter to avoid serious changes like >> this and UsersManager integration in the beta stage. >> >> Cheers, >> -- Pavel >> >> On Sat, Mar 2, 2013 at 11:26 AM, Pavel Krivanek >> <[email protected]> wrote: >>> This update broke some Pharo Kernel jobs. I will look at it later. >>> >>> -- Pavel >>> >>> On Fri, Mar 1, 2013 at 3:16 PM, Esteban Lorenzano <[email protected]> >>> wrote: >>>> 20578 >>>> ----- >>>> >>>> Issue 7617: make NBExternalEnumeration a subclass of SharedPool again >>>> https://code.google.com/p/pharo/issues/detail?id=7617 >>>> >>>> Issue 7605: moving a method into an extension does not notify the >>>> monticello package >>>> https://code.google.com/p/pharo/issues/detail?id=7605 >>>> >>>> Issue 7590: Kill MC History Data (>8MB of the image...) >>>> https://code.google.com/p/pharo/issues/detail?id=7590 >>>> >>>> Issue 7553: tons [delaySemaphore wait] process alive in image >>>> https://code.google.com/p/pharo/issues/detail?id=7553 >>>> (this one for check if it is fixed) >>>> >>>> Diff information: >>>> http://ss3.gemstone.com/ss/Pharo20/NativeBoost-Core-EstebanLorenzano.116.diff >>>> http://ss3.gemstone.com/ss/Pharo20/Kernel-EstebanLorenzano.1324.diff >>>> >>>> >> > >
