Raul Miller <[EMAIL PROTECTED]> writes: > Andreas Degert <[EMAIL PROTECTED]> wrote: > > If I just install one package, it's ok that a lot of output is > > produced, and maybe some questions are asked. But when I upgrade > > a batch of packages or install a new system, it's not an optimal > > solution. > > This doesn't scale to multiple machines.
That's right, asking questions interactively doesn't scale to multiple machine installs. The query service solves this (and my proposal to divide the install process into an interactive and a non-interactive part doesn't interfere), but answering the questions of the install scripts is the easy part of the game. Some random thoughts about scaling: There are a lot of packages with config files that can't be generated using some simple questions (on my systems packages like autofs, samba, lprng, X11, apt, dosemu, svgatextmode, netstd, pcmcia-cs, sane, samba, sudo, tetex, wine). This would have to be solved too if you want it to scale. If you want upgrades to scale, you have also to answer the old/new-configfile question by the query service (and the answer is only valid for one upgrade, for the next upgrade the choice might be different), and there must be a possibility to supply the merged configfile. When the machines have a similar but not identical configuration, for example hardware differences, answers must be supplied from a general, and a machine-specific database. ciao Andreas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
