Hi,
I'd like to define an PvmProcessDefinition using a fluent-API and have there an activity with sub-activities. Sub-activities are not connected to each other by means of transitions. They simply represent independent tasks to be executed.
For this composite activity, I'd like to define a following execution semantics:
- when this activity is entered via a transition, it starts all of its sub-activities in parallel
- when execution of all (or a certain number) of sub-activities has finished, the composite activity is considered to be finished as well and proceeds as usual using its outgoing transitions
Questions:
How something like this can be done using PVM APIs?
Are there any examples that can be used as a basis?
I'm very interested to see if I need to use createExecution for starting sub-activities and do all the book-keeping on my own. Specifically, how does one implement waiting for the termination of sub-activities? May be it is possible to receive some signals upon termination of each sub-activity?
Is it possible to have non-blocking waiting for termination of sub-activities?
Thanks in advance,
Leo