No bug in autoDrop

[Mirjana Majdandzic]

We finally managed to solve the problem of non-dropped activities while
autoDrop is set (which caused memory leak) in our dynamically scheduled model.

What happend was that  actions that were occuring at the same timestep
were treated as an Actiongroup. Each timestep, only the actionGroup as a
whole was dropped but not every single action it contained. So we changed the
type of the actionGroup used in the schedule by using SetConcurrentGroupType,
so that actions were automatically dropped.
And indeed the memory leak was gone.

Afterwards, I discovered in swarm-support that Doug has met the same
problem a few months ago (subject: Concurrent Schedule). I'm sorry that I
didn't read
that mail before and that I didn't make my problem clear enough.

I've three more notes:
1. In the documentation, in Schedule.html, an example of concurrentGroup is
mentioned. There may be a  mistake in the line:
 [concurrentGroupType setDefaultOrder: Random];
I think that "Random" should be "Randomized", because that 's how it's
documented in CompoundAction.html and ActionPlan.html (and Random is not
recognized as a valid identifier).

2. Another thing from the example that didn't work was:
 concurrentGroupType = [ActionGroup customizeBegin: [self getZone]];
Where "ActionGroup" should be "ConcurrentGroup" to let it work.
Does anybody know why?

3. Furthermore we still wonder why
 printf("autoDrop = %d\n", [modelSchedule getAutoDrop]);
returns 0 if you have set it to 1.

Bye,

-- 
Mirjana Majdandzic

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.
                  ==================================