On Mon, Aug 26, 2013 at 02:49:00PM -0400, Bill Nottingham wrote: > John C. Peterson (jcp@xxxxxxxxxx) said: > > On Sat, Aug 24, 2013 at 03:17:39PM +0800, Christopher Meng wrote: > > > On Sat, Aug 17, 2013 at 1:31 PM, John C. Peterson <jcp at eskimo.com> wrote: > > > > I would like to edit comps.xml to add a new package group for the tools > > > > that have already been packaged by the Formal Methods SIG. > > > > > > I just want to know if Group tag is needed or not,I never add Group tag in > > > any specs of mine. > > > > I was referring to the groups defined in comps.xml which is used by the > > installer during during the software selection phase of installation. > > > With respect to your particular definition of a comps group, you're creating > a large group with almost no default packages and a bunch of optional ones. In my mind, the primary issued being addressed by adding the new group is that these packages are currently in a group where *most* end users would not expect to find them (Engineering and Scientific). That group also has *many* packages in it, so some users may not notice them if they are not paying attention. > This sort of group definition is not the most useful, given that in the > installer (and with groupinstall) you'd only get the very small set of > packages (which may or may not be useful in isolation), and in other > post-install package tools, you'd be picking optional packages one at a > time. FWIW, *all* of the packages I listed that are already listed in the "Engineering and Scientific" group are currently optional packages. > How do you expect people to install these packages, and in what > combinations? For those user's who are knowledgable in Formal Methods, they will likely know what they want and what they don't need. For those users who are unfamiliar with such tools, they get a few packages to explore without a lot of duplication of basic capability (e.g. there are multiple SAT solvers). -- John C. Peterson, KD6EKQ mailto:jcp@xxxxxxxxxx San Diego, CA U.S.A -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
