Thread: [Cpu-users] Group Changes and CPU 1.4.1-RC1
Brought to you by:
matheny
|
From: Blake M. <bma...@pu...> - 2003-09-19 19:39:08
|
CPU 1.4.1-RC1 is available at: http://cpu.sourceforge.net/cpu-1.4.1-RC1.tar.gz and http://cpu.sourceforge.net/cpu-1.4.1-RC1.tar.bz2 There have been several major changes made in terms of the behavior of cpu. These include: - Refuse to add or modify users whose supplementary groups (or supplied primary groups) do not exist. This is the same behavior as the GNU shadowutils - When a user is deleted they are now automatically removed from any secondary groups they were in (same as shadow-utils). - When a users secondary groups are modified, they are removed from all old secondary groups and added to new secondary groups (same as shadow-utils). - When a users primary group is changed they are no longer added to that group via a memberUid (same as shadow-utils). - Refuse to perform a groupdel if the group is a users primary group. This is the default behavior of the shadow-utils. However, the shadow-utils allow the group id to be changed even if it is a users primary group. I think this is wrong, but cpu does the same thing. Should it? - Allow violation of POSIX naming conventions (for use with Samba) by specifying the -o switch I would appreciate feedback if people think any of the above behaviors are incorrect. The other thing I have been thinking about, is what the behavior should be when a user is added and -g is not explicitly set. Should a group be generated with the same name as the user, and make the users primary group that? Should there be a setting in the configuration file that specifies a default group? As always, feedback (and bug-reports) are appreciated. -Blake -- Blake Matheny "... one of the main causes of the fall of the bma...@pu... Roman Empire was that, lacking zero, they had http://www.mkfifo.net no way to indicate successful termination of http://ovmj.org/GNUnet/ their C programs." --Robert Firth |
|
From: Blake M. <bma...@pu...> - 2003-09-19 20:22:02
|
CPU 1.4.1-RC1 is available at: http://cpu.sourceforge.net/cpu-1.4.1-RC1.tar.gz and http://cpu.sourceforge.net/cpu-1.4.1-RC1.tar.bz2 There have been several major changes made in terms of the behavior of cpu. These include: - Refuse to add or modify users whose supplementary groups (or supplied primary groups) do not exist. This is the same behavior as the GNU shadowutils - When a user is deleted they are now automatically removed from any secondary groups they were in (same as shadow-utils). - When a users secondary groups are modified, they are removed from all old secondary groups and added to new secondary groups (same as shadow-utils). - When a users primary group is changed they are no longer added to that group via a memberUid (same as shadow-utils). - Refuse to perform a groupdel if the group is a users primary group. This is the default behavior of the shadow-utils. However, the shadow-utils allow the group id to be changed even if it is a users primary group. I think this is wrong, but cpu does the same thing. Should it? - Allow violation of POSIX naming conventions (for use with Samba) by specifying the -o switch I would appreciate feedback if people think any of the above behaviors are incorrect. The other thing I have been thinking about, is what the behavior should be when a user is added and -g is not explicitly set. Should a group be generated with the same name as the user, and make the users primary group that? Should there be a setting in the configuration file that specifies a default group? As always, feedback (and bug-reports) are appreciated. -Blake -- Blake Matheny "... one of the main causes of the fall of the bma...@pu... Roman Empire was that, lacking zero, they had http://www.mkfifo.net no way to indicate successful termination of http://ovmj.org/GNUnet/ their C programs." --Robert Firth |
|
From: Lukas K. <lu...@kn...> - 2003-09-19 20:43:37
|
Am Fr, 2003-09-19 um 13.52 schrieb Blake Matheny:
Hi!
Congrats in making cpu behave more like it's GNU equivalents. I tried to
test it right away, but somehow my testing-slapd died away (I'll try
again later).
> - Allow violation of POSIX naming conventions (for use with Samba) by
> specifying the -o switch
Thank you for this, again :)
> I would appreciate feedback if people think any of the above behaviors are
> incorrect. The other thing I have been thinking about, is what the behavior
> should be when a user is added and -g is not explicitly set. Should a group be
> generated with the same name as the user, and make the users primary group
> that? Should there be a setting in the configuration file that specifies a
> default group?
In Debian, there is /etc/adduser.conf, which contains
--- snip ---
# The USERGROUPS variable can be either "yes" or "no". If "yes" each
# created user will be given their own group to use as a default, and
# their home directories will be g+s. If "no", each created user will
# be placed in the group whose gid is USERS_GID (see below).
USERGROUPS=yes
# If USERGROUPS is "no", then USERS_GID should be the GID of the group
# `users' (or the equivalent group) on your system.
USERS_GID=100
--- snap ---
I would suggest to do it in a similar manner, so that the user can make
the choice. For example, AFAIR the default on SuSE-Linux is to have a
group 'users' set to every user's primary group, while the default in
Debian is to give each user his own group with uid=gid.
> As always, feedback (and bug-reports) are appreciated.
>
> -Blake
--
bye
Lukas
|
|
From: Blake M. <bma...@pu...> - 2003-09-20 04:45:02
|
This is implemented in CVS. Thanks for the suggestion. I'm in the process of writing some regression tests with DejaGnu which will likely be the last bit to go into 1.4.1 unless there are bug reports. -Blake Whatchu talkin' 'bout, Willis? > > ...snip... > > In Debian, there is /etc/adduser.conf, which contains > > --- snip --- > # The USERGROUPS variable can be either "yes" or "no". If "yes" each > # created user will be given their own group to use as a default, and > # their home directories will be g+s. If "no", each created user will > # be placed in the group whose gid is USERS_GID (see below). > USERGROUPS=yes > > # If USERGROUPS is "no", then USERS_GID should be the GID of the group > # `users' (or the equivalent group) on your system. > USERS_GID=100 > --- snap --- > > I would suggest to do it in a similar manner, so that the user can make > the choice. For example, AFAIR the default on SuSE-Linux is to have a > group 'users' set to every user's primary group, while the default in > Debian is to give each user his own group with uid=gid. > > ...snip... > -- Blake Matheny "... one of the main causes of the fall of the bma...@pu... Roman Empire was that, lacking zero, they had http://www.mkfifo.net no way to indicate successful termination of http://ovmj.org/GNUnet/ their C programs." --Robert Firth |