I agree. This is what I'm testing currently. Actually only the private methods which use the maps need to be synchronized.
But since I wasn't able to reproduce the problem and I have no time now to write a new multithreaded unit test for PageWindowManager I'm not 100% certain that the issue will be solved. Even DiskDataStoreTest doesn't reproduce it... On Wed, May 30, 2012 at 12:40 PM, Andrea Del Bene <adelb...@ciseonweb.it> wrote: > I don't wanna be too simplistic or too optimistic ( :-) ), but it should be > enough to make 'synchronized' the public methods of PageWindowManager...This > shouldn't cause performance issues since we have only few (two?) threads per > user and is better then building a complex error-prone implementation of a > synchronized bi-directional map. > >> Yes, could be. >> We need some sort of synchronization in this code for sure. >> >>> > > > -- Martin Grigorov jWeekend Training, Consulting, Development http://jWeekend.com