Passwd/Group changes status v3.
Philippe Troin:
> Remains:
> o Majordomo problem:
> Still waiting for someone to confirm that the following change is
> harmless. Otherwise, rex will roll out with the bug:
> /etc/passwd:
> majordom:*:30:30:majordomo:/var/majordomo:/bin/sh
> /etc/group:
> majordom:*:31:majordom
> To become:
> /etc/passwd:
> majordom:*:30:31:majordomo:/usr/lib/majordomo:/bin/sh
> /etc/group:
> majordom:*:31:
I've looked at this, and it seems odd to me. The majordomo package
contains indeed files with uid 30 and gid 31, but this seems to me to
be a mistake.
I'd be inclined to say that we should change the majordomo gid to 30
and offer the user the option (if they have majordomo installed) of
reallocating its files - either not at all, or by running find on
/var/lib/majordomo, or by running find on all local filesystems.
The majordomo package needs to be fixed.
> o Qmail controversy:
> What do you think Christian about using ``adduser -uid 6500{0,1,2,3,4,5,6}'' ?
> What should we do if adduser fails ?
The best would be for adduser to have an option to test whether
certain uids and names are allocated. Qmail could first say
adduser --check-free --uid 65001 qmail<spong>
adduser --check-free --uid 65002 qmail<wibble>
...
and then
adduser --uid 65001 qmail<spong>
adduser --uid 65002 qmail<wibble>
...
Use of set -e would prevent the package from installing if one of the
uids or names was in use with another name or uid, and adduser could
print a useful message.
> o Gnats doubt:
> Brian C. White <bcwhite@verisim.com> wants to have its current UID for gnats to
> be changed so that it matches a newly allocated GID for gnats.
> That raises the same problem that my previous proposal to roll out rex with stat
> ic uids in the 0-99 range for Qmail and change them later on with bo.
Why don't we give gnats its existing UID as a GID instead ?
Ian.
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-REQUEST@lists.debian.org . Trouble? e-mail to Bruce@Pixar.com
Reply to: