Quoting Ben DJ (bendj095124367913213...@gmail.com): > But, as you're apparently a Debian developer, and we're not, it seems > I'm 'being told' ? > > If that's the case, then we'll have to live without Debian Guest deployments.
On that part, there's probably something you're missing. When compared to many other projects you mentioned, Debian is solely and only based on volunteer work, which often resumes to "the work is done when someone needs it enough to push it and get it done". In the case of the feature you're talking about, I think that the need for the feature can well be understood (I actually myself have servers, at work, that deliver CentOS kickstart scripts by NFS...) and a first good star would be getting someone (beign a DD or not is really not important) to work on the implementation. Still, as Frans said, it has a "cost": one more module in the installer's kernel, more code in the network-preseed package, adding an NFS mounting utility....all this going in the installer's initrd....that has important constraints on size as we're very conservative about memory usage (contrary to other projects you mentioned, Debian is also widely used in environments where resources are highly constrained). So, even if the feature is implemented, we would have to ponder about its impact. However, things never came up close enough to a working implementation for the real impact to be estimated, as far as I can see. So, back to your remark: there is no decision instance, in Debian, to take a "supreme" decision and say to our users (who are always seen as potential contributors) "OK, you're told, there will never be NFS supprot for preseeding"...and I don't really doubt that someone coming with a fully-working implementation would have string arguments in a discussion. After all, this is mostly how the graphical installer project started a few years ago..:-)
signature.asc
Description: Digital signature
