On 2012-3-20 06:05 , Ryan Schmidt wrote: > > On Mar 19, 2012, at 13:50, Joshua Root wrote: > >>> It might be useful to have a "default_subport" directive. When a user tries >>> to install a port that has subports, if default_subport is set, MacPorts >>> installs that subport instead. Otherwise, MacPorts tells the user to select >>> one of the subports to install. >> >> Let's try to solve the problems with appropriate mechanisms, not just >> bash all the non-nails with our subport hammer. > > It sounds like you think something I said was completely crazy, and I'm not > sure what it is; I thought my suggestions were rather good and sane. What was > it that you objected to? How would you approach the problem instead?
It's not completely crazy, it's just not addressing the underlying problem, which is that we want features that would be best provided by objects that do not behave like ports. What you want is a collective label that you can use to refer to a group of ports; that label should not itself be a port. We want a record that a port has been replaced by another; that record should not itself be a port. We want aliases that allow you to refer to a port by different name; those aliases should not themselves be ports. - Josh _______________________________________________ macports-dev mailing list [email protected] http://lists.macosforge.org/mailman/listinfo.cgi/macports-dev
