>> Now you can say that this is not your use-case, but Metacello >> configurations are not my use-case either. > > I say nothing. I just totally convinced that if we do not want to stay an > obscure little toy we have to capture > dependencies and scripts do not capture that.
I totally agree. > For me metacello is not the panacea but it exists and if we would spend a > little time then the system would benefit from it. > Now there are problems because I have to fix configurations for systems I > have no clue how they are built. This makes me wonder why you feel obliged to fix these problems? > So what your attitude just create is more burden on my shoulder. Sadly I > would love that we do not have to use your projects > but we cannot so I should do the fucking stupid job of extracting the > dependencies in your project. Grease, Seaside, Magritte, Pier, ... have very well statically defined dependencies declared in each package (extension methods in GRPackage). There are numerous tests that ensure consistency (they are verified with each build). Dale has written tests that verify that the Grease declarations and Metacello definitions are in sync (don't know if they pass, they are not run if Metacello is not in the image). Furthermore, I have repeatedly posted dependencies graphs of various parts of the system. Also Jannik has repeatedly produced matrices with dependency information (I could never understand them). Either way, I thought we were are all fluent in one or the other tool to tell us outright about broken dependencies? > It would cost you 3 min and it costs me hours but apparently your are right > and I'm wrong and this is my stone. > So I will continue. How? I do not even know what is wrong. Lukas -- Lukas Renggli www.lukas-renggli.ch
