tag 593561 + confirmed tag 593561 + pending severity 593561 important thanks
On Thu, 19 Aug 2010 19:34:23 +1000 "Trent W. Buck" <[email protected]> wrote: > Consider the case where multistrap.conf contains: > > [General] > debootstrap=Y > aptsources=X > [...] > The user > > 1. runs multistrap to succesful completion > 2. realizes that she forgot to include "mg" > 3. adds "mg" to packages= line in multistrap.conf > 4. re-runs multistrap > > At this point, multistrap will be attempting to bootstrap while > sources.d contains BOTH multistrap-X.conf (from the first run) and > multistrap-Y.conf (from the second run). That sounds like a bug. If a source is not included in the bootstrap option, it should not be accessible to the bootstrap operation, only the subsequent apt sources. > If Y is a file: repo and X is an http repo, this can make the second > run take longer / download more than it needs to. > > To work around this, I run the following prior to each multistrap run. > If this doesn't conflict with the "stateless" approach, it could be > done in multistrap. > > rm -f $dir/etc/apt/sources.d/multistrap-*.conf Multistrap should do that itself - indeed it should probably extend that to rm -f $dir/etc/apt/sources.d/* It already tries to do this: unlink ("${dir}etc/apt/sources.list.d/multistrap.sources.list") if (-f "${dir}etc/apt/sources.list.d/multistrap.sources.list"); unlink ("${dir}etc/apt/sources.list") if (-f "${dir}etc/apt/sources.list"); I'll extend that so that ${dir}etc/apt/sources.list.d/ is empty at the start of each multistrap run - stateless doesn't mean that multistrap can afford to ignore / undermine it's own configuration settings. -- Neil Williams ============= http://www.data-freedom.org/ http://www.linux.codehelp.co.uk/ http://e-mail.is-not-s.ms/
pgpwrvUKPFGrX.pgp
Description: PGP signature
