I discovered that restarting Server service does in fact RESTORE the share!
The problem with restarting this service using a scheduled (after boot) script, is that this is one of the most important services of Windows.
In my SBS7, trying to stop Server service, means you have to also stop (and then take care to re-start),
Microsoft Exchange System Attendant (oh)
Microsoft Exchange Information Store (yikes)
DFS Namespace (!)
...not cool. BUT IT WORKS! After I restart those services, automagically my merge volume is again shared!
But worse of all? After reboot, Liquesce died on me. Service starts and then stops and I get various events (this is for the Issue Tracker, I will not give details here)...
(after second reboot it worked but really took its time)
The thing is that there is another OPEN SOURCE project that has dealt AND solved the exact same problem. It is the same project that initially needed the server service restart: TrueCrypt.
In TrueCrypt, a (virtual) volume can be marked as "favourite" and this indeed allows for it to get the same letter, be reshared etc.
In fact even better, there are even "System Favorite Volumes" that get mounted early enough to be able to be used by the system itself! EXACTLY WHAT I WOULD LOVE TO HAVE.
In fact since TrueCrypt involves, erm... encryption, in our case things should be simpler!
So can this be communicated to Dokan team, or maybe you could check out TrueCrypt's part of the code that deals with this?