Initially this was planned to be done way after many other improvements. However one of friends kindly reminded me that real undo / redo capabilities were essential to any modern application.

Obviously the real-time multi-user structure of StakePoint database architecture is not really in line with precise undo operations. Any undo / redo action can only be tentative. No success guaranteed as any user can do something which is in conflict with the temptation to undo.

For the moment the undo mechanisms are written in a local file which is scratched at start-up. We can easily modify this later, either in an appended file spanning multiple sessions, or by storing the total history of undo within the database itself, making it accessible to reconstruct the system. For the moment we decided to keep things simple. It is anyway important that we keep the positive approach against data integrity. A safe verified attempt is better then a simple unverified modification.

All complicated undo / redo actions are now sort of covered, and we are eager to release this with next release ASAP.