On Mon, 2007-04-09 at 13:57 -0400, Christopher Aillon wrote: > Matthew Miller wrote: > > Speaking from > > experience as someone who rebuilds a lot of packages, this is exactly the > > reason to rebuild all packages at some point in the development cycle. It > > sucks to discover that a package doesn't actually rebuild cleanly anymore > > when you really need to do it urgently. > > It does suck. Firefox security updates in Fedora have been delayed due > to compiler and other toolchain issues, due to GTK+ changes, due to > pango changes, etc. > > Still, That does not warrant a rebuild at this point because new > features can crop up and we lose testing on a stable set of packages. > > I'd rather see a tinderbox-like build happen for these things to catch > them. Rebuilds should happen as quickly as we can churn them out, but > these builds should be _throwaway_ builds. They should not be released. > The main purpose should be to catch things that no longer rebuild > properly. This will solve the issue without needlessly pushing new > packages to people over a cosmetic change this late in the game. I do this for the stuff I maintain using scratch builds in brew. This could easily be expanded to an X-line script that just always runs and walks the CVS tree building whatever is the latest tag for devel for a package. It might also be useful to have this kind of system be opt-in for package maintainers rather than dumping everything in at once. Going even further, we could also opt-in spare boxes we have as builders for this system. -- David Cantrell <dcantrell@xxxxxxxxxx> Red Hat / Westford, MA
Attachment:
signature.asc
Description: This is a digitally signed message part
-- 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
