Adding workstation to a samba pre 3.0 controlled domain

Gerald Carter gcarter at
Wed Nov 8 14:30:34 GMT 2000

Simo Sorce wrote:
> Tonight I took a look in the code both from samba HEAD 
> and samba TNG. Samedit code is different from the HEAD 
> one, but while I prefer the HEAD one I think both 

samedit in HEAD is broken I think.  

> are not good. I think a new function should be 
> added in passdb_ops structure:
> bool new_user_entry (char *user_name, int local_flags, char
> *new_passwd);

No need for a new function.  I'm working on a 
redesign for the passdb API.  A lot can be simplified.  I'll
post an RFC later today.

> The procedure should be implemented in various backends.
> Also user creation in system passwd should be demanded 
> to the backends. This would easly permit to change the 
> way user are created on the system. It comes in mind 
> the possibility to use winbindd on the Samba DC or LDAP
> for both samba and system users, the way user creation 
> is performed just now will prevent this thing.
> Obviously a convenience procedure may be set in 
> passdb.c for backends that will not support the function.

I looked at the winbind code yesterday.  Tim, it looks like
it deals is user information formatted for /etc/passwd
entries.  This makes it not flexible enough to use as a buffer
between Samba and the passdb backend.  Am I correct?  Am I 
making sense?  I'll follow up on this later (with a lengthy

Cheers, jerry
   /\  Gerald (Jerry) Carter                     Professional Services
 \/  VA Linux Systems   gcarter at       SAMBA Team          jerry at                     jerry at

       "...a hundred billion castaways looking for a home."
                                - Sting "Message in a Bottle" ( 1979 )

More information about the samba-technical mailing list