On Sun, Jul 10, 2011 at 11:59:36PM +0100, Peter Robinson wrote:
On Sun, Jul 10, 2011 at 11:49 PM, Matthew Garrett
<mjg59(a)srcf.ucam.org>wrote:
> Finding that something that didn't build, hasn't been changed and still
> doesn't build isn't terribly random :) We could do better, but
> introducing more mass rebuilds as part of the release cycle isn't going
> to make things significantly better when we're already failing to deal
> with the fallout from the mass rebuilds we *do* carry out. Fixing this
> is a process issue rather than one that's fixed by having FESCO mandate
> a mass rebuild after every change that could conceivably cause breakage.
>
>
If a mass rebuild causes breakage its a problem which ever way you look at
it, the package will eventually be rebuilt and cause the breakage, in my
opinion your better off doing that in a controlled manner rather than
sweeping it under the carpet and hoping no one will notice.
I agree, but if we're failing to deal with FTBFS with the number of mass
rebuilds we currently have, introducing more mass rebuilds doesn't
magically make things better.
--
Matthew Garrett | mjg59(a)srcf.ucam.org