i don't mind, but sync with Alejandro and Thomas.
and make a very clear note about it in the release notes. as this is a backwards incompatible change. (i agree: a reasonable one)
Okay, I committed the change, the test suite still runs fine.
The comment in the release notes is a good thing, where I can add something to it?
*Copied from post in mailing list*
Are we considering the use case of "having non ended subprocesses, when
parent process is ended"?
With this solution, this would throw an exception, right?
"that you should not signal a process instrance that has been ended
sounds trivial to me" - Tom
well, that's what _jbpm itself_ does, when returning from subprocesses
to ended parent process.
You still can have "non ended subprocesses, when the parent process is ended". Because of the additional fix proposed in https://jira.jboss.org/jira/browse/JBPM-1471 this is no problem, since an ended sub process only signals the super process if it is not ended.
This fix was necessary anyway, otherwise canceling a process instance (which cancels sub process instance, which notifies the parent) wouldn't have been possible any more.
This is very good.
I somehow assumed (from comments), that the proposed fix was not accepted.