On Wed, May 16, 2007 at 03:04:08AM +0530, Rahul Sundaram wrote: > Patrice Dumas wrote: > > >>Have you looked at how many major regressions we have > >>for every release? > > > >Not precisely. > > Please do. This is a QA process and to understand the need for it you > need to know what drives it. Spend sometime reading through end users > forums. The question is: how do these regression distribute among Core and Extras (based on weight like number of packages in repo and also number of packages used by users of course). I'm under the impression that the regressions that made most of the noise were part of Core and that already had the QA systems being discussed in place (updates-testing, mandatory announcements etc). And OTOH we decide that packages don't need rebuilds pushing any broken package out there to break in users' hands instead of during the development cycle and OTOH we raise burocratic hurdles w/o a real gain. Ask yourself: Will an announcement of a package update really increate package *quality*, e.g. add to a better QA? Not really. Don't understand me wrong: I like package update announcements. But they will not magically increase package quality and reduce regressions. Same for an unused updates-testing repo. A better QA? Force package maintainers to push their packages into updates-testing and have them enable updates-testing *on their own systems*. That way you get a couple of hundreds skilled people actually using updates-testing ... -- Axel.Thimm at ATrpms.net
Attachment:
pgp34Es6acoql.pgp
Description: PGP signature
-- Fedora-maintainers mailing list Fedora-maintainers@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers
-- Fedora-maintainers-readonly mailing list Fedora-maintainers-readonly@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers-readonly
