Ricardo Newbery wrote at 2008-1-30 12:24 -0800:
... Has anyone given thought to maybe providing a more general class with conflict resolution in the core distribution?
There can be no *general* class with conflict resolution. Conflicts usually are fatal. Precise knowledge about the intended use is necessary to determine whether and how a conflict can be resolved. Take a look at the "Length" class. It resolves all conflicts in a way compatible with maintaining the length of a data structure. However, you must *NOT* use it to assign unique integer tickets as it may return the same value in different threads. Without conflict resolution you would get a conflict when different threads take a ticket concurrently; with the "Lenght" conflict resolution, the tickets are no longer unique. -- Dieter