Re: [ANNOUNCEMENT] TEST RELEASE: Cygwin 1.7.33-0.6

[Corinna Vinschen <corinna-cygwin at cygwin dot com>]

On Nov  7 07:37, Christian Franke wrote:
> Corinna Vinschen wrote:
> >On Nov  6 21:06, Corinna Vinschen wrote:
> >>On Nov  6 20:51, Christian Franke wrote:
> >>>Corinna Vinschen wrote:
> >>>>On Nov  6 19:34, Christian Franke wrote:
> >>>>>But why does
> >>>>>   mkpasswd -l (no host) -- adds a prefix
> >>>>>   mkpasswd -l THISHOST -- does not add a prefix
> >>>>>when the machine is in a domain? Not consistent, IMO.
> >>>>That's right.  The reason is that the machine name is treated as a
> >>>>foreign machine.  In theory, this should always generate names
> >>>>with prefixed machine name, but this is an entirely different
> >>>>code path in mkpasswd/mkgroup.  I guess this should be fixed.
> >>>>
> >>>>I wouldn't be unhappy about help...
> >>>I would only fix it back to the old behaviour (mkpasswd -l = no prefix),
> >>>sorry :-)
> >>>
> >>>At my real job we run several build & test machines which are members of a
> >>>domain but use various local test user accounts (with no collision with
> >>>domain users due to name space rules). Loosing the ability to use
> >>>prefix-less local user names would break various existing test scripts
> >>>(which are also used on Linux).
> >>>
> >>>Generated emails would have a from address with HOST+USER name part which
> >>>might give interesting results if the mail system somehow interprets the
> >>>NAME+EXTENSION address syntax...
> >>>
> >>>So there are use cases where prefix-less local user names are needed. This
> >>>should be still supported, e.g. by mkpasswd -l, IMO.
> >>But then... why not keep mkpasswd -L and use that instead?
> >On second thought, it's completely wrong to allow printing local
> >accounts from another machine without prefix.
> 
> I agree.
> 
> >In theory there should be only one option -l [machine], which prints the
> >local accounts of the current machine unprefixed (standalone machine) or
> >prefixed (domain machine), and always prefixed for a foreign machine.
> >The -L option can just go away.
> 
> I disgree.
> 
> Why not keep the old behavior of -l/-L for user names of current machine for
> those uses cases which rely on it?

You are always free to change the passwd/group files manually:  

  $ mkpasswd -l | sed -e 's/^[^:]*+//' > /etc/passwd

Does a system tool really have to generate a scheme which might lead
to collisions?  The Administrator account is only one of them:

  $ mkpasswd -l -d | sed -e 's/^[^:]*+//' | grep Administrator:
  Administrator:*:197108:197121:U-LOCAL_MACHINE\Administrator,[...]
  Administrator:*:1049076:1049089:U-DOMAIN\Administrator,[...]

> Those users who are happy with prefixed
> local user names and non-prefixed domain user names would simply no longer
> need to use mkpasswd (which is good).
> 
> Package search shows 156 usr/bin/*-config scripts. How many of these use
> mkpasswd?

I don't know, but I'm trying to fix up at least the ones I have under
control.  The other will follow over time.

> BTW: None of my Linux machines have local user names with own HOSTNAME as
> prefix :-)

Your Linux machines usually don't have to maintain user and group accounts
from various account DBs.  This is all nice and easy in a single domain
scenario, but as soon as you maintain multiple domains, you will have
name collisions.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

Attachment: pgpkpUPcJYK6v.pgp
Description: PGP signature