On Wed, Jul 9, 2008 at 10:26 AM, Josh Boyer <jwboyer@xxxxxxxxx> wrote: > I'll argue that not every package upgrade is worth a Feature > designation. But the major ones should be. Firefox 3 had one. I > believe OpenOffice.org should have one. For a major rpm upgrade, there > should be one as well. Distill that feeling into a generally applicable statement that package maintainers can use as a conditional test as to whether or not they should file a feature. For example... the python-matplotlib that I just introduced into rawhide.. it has enhancements, and some internal api changes. Does it matter enough as a package to be a Fedora Feature? If I can get a new gdl built with the python module support enable..will that matter? I need a bright line test that makes sense that limits how much of my individual judgement has to be used for the packages that I maintain. I need a checklist or a scorecard....I need a Cosmo quiz. And its completely okay if the test/checklist/scorecard that we think is best can be fully created yet. For example.. if we feel that a packages "popularity" is important to rank something as a feature, then we put that on the checklist and we then go out and build whatever popularity metric is necessary to support that part of the checklist. -jef"is definitely not in favor of any popularity metric"spaleta -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
