Re: force-confnew (was: Document correct buildd chroot setup somewhere?)
- To: Philipp Kern <pkern@debian.org>
- Cc: debian-wb-team@lists.debian.org
- Subject: Re: force-confnew (was: Document correct buildd chroot setup somewhere?)
- From: Wouter Verhelst <wouter@debian.org>
- Date: Thu, 6 May 2010 14:44:01 +0200
- Message-id: <[🔎] 20100506124401.GK3828@celtic.nixsys.be>
- In-reply-to: <20100405183616.GA20708@kelgar.0x539.de>
- References: <201004051131.11557.sf@debian.org> <20100405101812.GA18055@kelgar.0x539.de> <20100405111635.GR19557@mails.so.argh.org> <20100405115421.GA5236@roeckx.be> <20100405115831.GS19557@mails.so.argh.org> <20100405121838.GA5417@roeckx.be> <20100405122349.GT19557@mails.so.argh.org> <20100405124740.GA5591@roeckx.be> <20100405183616.GA20708@kelgar.0x539.de>
On Mon, Apr 05, 2010 at 08:36:16PM +0200, Philipp Kern wrote:
> This doesn't mean that we shouldn't fail on obvious bugs in a package.
> I'm just not sure if we should allow bystanding packages that happen
> to be pulled in somewhere (and not being uninstalled/purged) to screw up
> the build process *in a non-deterministic way*.
Except that it's not actually nondeterministic.
Missing build-conflicts are just latent packaging bugs; while we don't
guarantee that they don't exist, having the ability to detect them, even
as a byproduct of some other process, is a good thing, IMHO.
At any rate, there are many ways in which buildd/sbuild can fail. This
is just one of them; and that is why we have a human handling the
output, rather than make it an automated process.
--
The biometric identification system at the gates of the CIA headquarters
works because there's a guard with a large gun making sure no one is
trying to fool the system.
http://www.schneier.com/blog/archives/2009/01/biometrics.html
Reply to: