Tom Horsley wrote: > On Fri, 12 Dec 2008 13:40:00 +0000 > Anne Wilson wrote: > > >>> Fact: the push of dbus directly to stable was a mistake.. it was not >>> intentional >>> >> Devs and maintainers are not allowed to make mistakes! That would imply that >> they are human beings, and we can't have users with such dangerous thoughts! >> > > Well, that was what my original proposal was about. Pore old humans > can decide to push things to "almost updates", but only a computer > which verifies the updates can be applied successfuly to at least one > test instance of a fedora install can push the updates out of "almost > updates" and into "updates". > > Of course, I wasn't thinking about the dbus incident specifically, > since that version of dbus applied with no conflicts, it would have > made it through the automated test as well. I was thinking more > about all the conflicts which continually show up where some new > package needs a new library and the new library isn't available > yet, etc. Just a quick automated test to prevent obvious problems > from reaching updates, but short of a full-blown dependency analysis > of the entire repo. > > FWIW, the 3 layer model is used to great effect in everyday business. First, there's "testing" where the developers get to play to their hearts content and, hopefully, get a product to "production" level. Then the product goes to "qa" or "qc" where it "burns in" for awhile with other products that may or may not "play nice" with it (aren't double quotes wonderous little things!). If they don't "play nice" together then it goes back to testing for more work and then back to qa/qc until it all works as planned. Then, and only then, does it move to production. I understand that Fedora is a bunch of folks doing the work on a volunteer basis but that just makes the idea of a qa environment that much more useful. Kevin -- fedora-list mailing list fedora-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
