Zeeshan pointed me at this thread https://www.redhat.com/archives/libosinfo/2012-December/msg00174.html asking whether I thought it is better for the OsinfoInstallConfig class to have a 'config-params' property (as per that patch), or a more general 'install-script' property. Of course I thought it should have neither! IMHO the whole design is overly convoluted. The InstallScript class is creating a copy of the InstallConfig object instance, passing in the InstallScript so that the new InstallConfig object can have a 'config-params' set associated with it. This is so that the InstalLScript can then use that to apply datamap translations. This is all complete overkill - the InstallScript class already has info about the datamaps and can apply them to the original InstallConfig object instance it has, without needing to create a throwaway copy. This series reverts the last 5 patches in that quoted series and applies one simpler patch at the end which does the same job without needing to directly associate the InstallConfig & InstallScript classes at all. _______________________________________________ Libosinfo mailing list Libosinfo@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libosinfo