On Tue, Jun 30, 1998 at 12:54:16PM -0400, Adam P. Harris wrote: > > I don't know if this issue has been discussed, but I would like to > > know, if there will be support of pam (plugable authentication > > modules) in login. > > I believe we shall be putting PAM back in as our auth infrastructure > for Debian 2.1. AFAIK, we backed out of PAM in mid (?) 1997 because > it was not stable (?), but now all is well and not having PAM is a > pretty major technical flaw of Debian today. > > And we *hate* technical flaws. Yes we do! > However, PAM support is not listed as a goal in > http://fatman.mathematik.tu-muenchen.de/~schwarz/debian-roadmap/ . Isn't it? Everyone who's mentioned it has been saying slink. I figured then that it would happen in slink probably.. > Leader, etc: > > Even if we trash release goals as our criteria for doing a release, > which I think we should do, we *still* need goals. Just don't base > releasbility based on them. Here's my list: Certainly. > * FHS support This is an example of a goal that need not be met by all packages for release. > * PAM integration This afaik is only useful if most/all things use it.. This means login, shadow ... what else? This is a fairly big goal I suspect, but how big is big? > * better support for large-scale installations (i.e., diskless > clients, dataless clients, non-interactive installation/upgrade) > * consistent document infrastructure (metadata, subject scheme) These seem to be more long-term goals I think.. Certainly I wouldn't want to hold up a release for them, nice as they are for goals. Indeed, some of it can't happen now. Befire we can do thinks like have dpkg and apt just install things quietly in the background, it's gotta first offer a way to configure things another way than in the postinst. Linuxconf seems promising, but we can't use it yet to that degree. Hopefully modules will be able to be written in perl or C or whatever you want eventually and that'll be a non-issue.
Attachment:
pgp_8Pz_Vlqp9.pgp
Description: PGP signature