Hello Adam,
I'd like to correct something after our quick chat about {c,}debconf:
the former is used within d-i, but installation in /target is
performed using the latter; meaning keeping an eye from the d-i side
is nice, at least at this point of the release cycle.Colin Watson <cjwat...@debian.org> (10/12/2012): > I've just uploaded debconf 1.5.48, which fixes two regressions I > introduced several months ago (#686071 and #690755). I regret not > noticing these sooner. Both affect only dpkg-reconfigure: the first > is weird collateral damage to triggered packages which I can easily > forgive anyone but debconf hackers being completely stumped by, and > the second is just embarrassing because it breaks any use of > dpkg-reconfigure on more than one package. I can easily imagine at > least the former being the root cause of release-critical bugs > elsewhere. > > The other non-documentation change here is a change to debconf's GTK > frontend to do some additional sanity-check probing. This arises from a > bug that confused us for months during Ubuntu upgrade testing, where GTK > was temporarily unusable at some critical points during upgrade; this > kind of thing is far from uncommon regardless of the exact cause, so I > feel that extra checks in debconf are worthwhile for wheezy. Given I'd like to have a fix for #679327 (finally diagnosed!) in rc1, pushing the update from 1.5.46 to 1.5.48 into testing soon would be appreciated; that would clear up the path for a possible 1.5.49. Mraw, KiBi.
signature.asc
Description: Digital signature
