Joachim Breitner wrote: > Hi, > > Am Freitag, den 27.02.2009, 14:40 +0100 schrieb Steffen Moeller: >> Joachim Breitner wrote: >> [...] >>> He first changed every occurance of --yes with --yes --force-yes, and >>> changed every occurance of that back to --yes (configurable with a >>> variable). This included this change: >>> >>> - chroot $INST_DIR apt-get --yes --force-yes install pkg-fso-keyring >>> + chroot $INST_DIR apt-get $APT_OPTIONS install pkg-fso-keyring >>> >>> but in this case, the --force-yes was there originally and for a reason. >> Right. And I am sorry for this. >> When we have that pkg-fso-keyring package signed by a Debian developer, the >> --force-yes >> could go, right? > > No, you can’t sign individual packages, only repositories. I don’t think > that pkg-fso-keyring would be admitted to the main repository. > > A cleaner solution would be to include the pkg-fso key (or it’s > fingerprint) in install.sh and feed it to apt-key before installing from > the pkg-fso repository.
This sounds nice to me. Please be so kind to send me a signed email of yours of the key's fingerprint. > But didn’t we want to upload to main real soon anyways :-) :) but we don't want to be forced to do so sooner than we want to. >>> There were a few changes to the installer recently, and it seems they >>> were not completely tested. Not sure about the mount problem though, >>> without an error log. >> There was no run from scratch, which would have been good, but is somewhat >> tough because >> of the long run time....and since the /mnt/debian is not removed after the >> install, I >> would not have spotted that one. > > So is there a bug now left in the script? I am not sure. I personally like it the way it is, but it would possibly be cleaner to remove INST_DIR if it has not been existing before. I'll add that. Many greetings Steffen _______________________________________________ pkg-fso-maint mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/pkg-fso-maint
