Hi Michael. I'm glad you are running your client with ONLY the Magma-client package loaded as this hasn't been tested in long time.
As such, you have found a mis-categorized method. I recently changed #getNextOid to check to see whether it has run out of oids. In doing that, I made it use a method that is only part of the Magma server package.
My first inclination is to move #lastUserObjectOid into the Magma client package. But now I am re-evaluating the wisdom of this new error check in the first place. There are enough oids available to last for thousands of years under normal circumstances, so I just don't ever see it coming up, so why make the computer do that comparison over and over and over again?
Please revert #getNextOid by removing the first three lines and leaving it to just say:
^ nextOid _ nextOid + 1
This should clear up your DNU.
- Chris
--- Michael Roberts mike@mjr104.co.uk wrote:
Hi,
I've loaded Magma client (125) and server (105) from Monticello into separate 3.8 images and I get a DNU in the client image when I try and connect.
The client eventually ends up in MaOidManager>>getNextOid which sends #lastUserObjectOid to MaOidCalculator. This isn't defined in my client image.
However, it seems that it is extended in Magma server. Does this need to be pushed down into Magma client?
Also, is there a way of knowing which versions are considered stable? I just grabbed the latest versions but I would grab more 'blessed' versions if these numbers were indicated somewhere?
Thanks!
Mike