Re: why should packages migrate to testing automatically
From: Kevin Mark (kmark+debian-user_at_pipeline.com)
Date: Tue, 1 Mar 2005 23:10:28 -0500 To: firstname.lastname@example.org
On Tue, Mar 01, 2005 at 07:48:19PM -0700, Paul E Condon wrote:
> On Tue, Mar 01, 2005 at 06:36:22PM -0500, Kamaraju Kusumanchi wrote:
> > This might be nitpicking (sorry if it is), but I could not find the
> > answer in google.
> > I think it is a bad idea to allow packages to migrate to testing
> > automatically. Sometimes a maintainer might want the packages to stay
> > in unstable than in testing. A particular package might call for more
> > attention from Sid users to expose its bugs...
> > My personal opinion is that a package should migrate only upon the
> > consent of the maintainer.
> > Adv:-
> > 1) Testing will probably have less RC bugs than there are currently
> > have -> improved stability in testing -> less work for the bug
> > squashing party -> less release cycle.
> > 2) Maintainers make better choices and they know when the package is
> > ready for "testing".
> > 3) Gives the users of sid sufficient time to find a bug and report it.
> > Dis adv:-
> > 1) Updates to testing might be slower.
> > What do others think?
> > raju
> The rules for moving packages from unstable to testing are described in
> the debian developers reference (manual) section 5.13.2. I have not found
> a rational for this rule set, but my impression is that Debian tries to
> automate that which can be automated, but with controls and escapes.
> A maintainer has the option of selecting an urgency level of high,
> medium, or low, which hold the upload in unstable for 2, 5, or 10 days
> respectively. I presume that a maintainer can fold into his decision
> his own sense of how confident he is that the upload does not
> introduce a subtle bug. Thus he might downgrade the urgency of a bug
> fix if he feels maybe it needs some extra testing, even if it
> addressing a rather critical security problem. But maybe not. Maybe a
> bunch of broken systems is better than a bunch of compromised
> systems. But this is just my speculation. I don't know the rational
> that drove the Debian decision.
> Paul E Condon
I have done some research on the Debian process. One factor is the
urgency level as Paul said. But there is a Debian script that takes more
factors into consideration: Britney. I have not read the source for this
but it is available to inspect.
you can checkout my new diagram (I moved it to a new location) for any
comments or questions. It's not complete but I'm working on it.
-- counter.li.org #238656 -- goto counter.li.org and be counted! (__) (oo) /------\/ / | || * /\---/\ ~~ ~~ ...."Have you mooed today?"...
-- To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org with a subject of "unsubscribe". Trouble? Contact email@example.com
- application/pgp-signature attachment: Digital signature