Just like most "meta" packaging systems this looks like a major headache for anything else but software which has no real dependencies outside what itself provides. The example you give, Oracle, IBM products, etc, are relatively easy to install when compared to more "basic" stuff like, oh, KDE for example. The amount of dependencies and the algorithms to resolve them in a sane matter are not something to you're likely to implement properly into a tool which needs to support a whole lot of package formats in any other way than just basic functionality like bluntly installing and uninstalling a single rpm/deb/msi/whatever.
Anyway, it looks like you're trying to plug your own product. Nothing wrong with that, except that it's probably the first time most of us here have heard about it in the firstplace (it was for me anyway). If that is the case I estimate your chances to be slim to find much adoption for something which is both unproven and underdocumented. I don't have a solution for the dilemma this community seems to be in. If it were my call, I'd pick the packaging system which is very well documented, performs well and has a lot of packages for it already. From where I'm sitting that qualifies RPM and DPKG. When you throw in the requirement for the proper tools and documentation to handle those packages (install, remove, dependency resolving, etc) and to create them in the firstplace, I'd say DPKG/APT/devscripts wins hands down. But maybe that's just because I like Debian :-) This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
