When I compile with the control patch and use pbs* it never finds the parameters it needs in order to work. Not sure if this is something that is a bug with the control patch or with pbs. I will assume it is with ldap-control and the blame should be placed there. Obviously, I have found the workaround by compiling twice but it makes it rather hard when you are trying to keep everything consistent.
The system itself compiles fines with either control or without control so to answer your question, it is more fundamental than compile time. I would like to make a suggestion because I am getting a report of intermittent problems. Where customers are working fine and then all of a sudden they get the famous 5.5.3 error. The system is working correctly but every once in a while it does not get the information back from pbs and in turn rejects the relay. Now to my suggestion, can we make it run on a tcp port rather than udp? I know from a error standpoint you want to keep the system running but in the case of the client I would rather see the connect fail than send them a 5.5.3 error. (Cuts down on support calls we can not solve). I will tell you this, on the old system where we called the external script /usr/pop3-record I have it working on 4 mail servers with NFS and everyone was working correctly all the time. I had the cdb on the NFS mount with the compile calling the tmp on the local machines. This rocked. Never failed, worked every time. So back to my suggestion on the tcp, I think it would be beneficial to stop working if there is a problem then having 500 people call you with the same error. Just my two cents. Feel free to comment. Thank you for all the great work everyone has done to create qmail and the appropriate ldap patches. It makes our lives so much easier here. Gene Parks VIP Direct
