Hi Jerry, On Wed, Jul 27, 2005 at 08:47:34AM -0500, Gerald (Jerry) Carter wrote: > [EMAIL PROTECTED] wrote: > > | Fix amazing and long-standing bug where user-accounts > | are just crippled accounts (accounts without AcctCtrl set) > | after a vampire-process. > | > | New Accounts tend to hace no acb_info at all which means "0" > | (ACB_NORMAL). Unless 0 becomes not 0 we don't do anything > | and set *no* acctrl for normal users at all (!). Those > | crippled users now don't show up in usrmgr since > | 3.0.20somethings ldap-routines now finally test if > | the attribute is there. > > Guenther, > > You added the change to start looking at the separating > searches based on the ACB value for 3.0.11 I thought. > Or are we talking about a different change to ldapsam > now?
This is something completly different. What I've added a while ago was: when enumerating users, derive ldap_filter from the requested acb-info. But this one here is an important change to "net rpc vampire": vampire 100 users and they all had *no* acb_info set in ldap (!). Now the new search-semantics for enumerations in LDAP (since the early 3.0.20betas) simply skip accounts that have no acb_info at all (for good reason). 3.0.20 will trigger (even without my fix which just fixes a broken "net rpc vampire" for future use) a couple of mails like "hey, since I upgraded to 3.0.20 most of my users are no longer displayed in usrmgr". Maybe we have to provide a magic selfrepair of vampired accounts. Cheers, Guenther -- Günther Deschner GPG-ID: 8EE11688 Novell / SUSE LINUX [EMAIL PROTECTED] Samba Team [EMAIL PROTECTED]
pgp6iBBploHeC.pgp
Description: PGP signature