Add an issue in JIRA for it. Then this will be tracable...
Issue open: JBPM-455
By the way, same problem happens the other way around, when invoking a subprocess, if a read variable has not been set, the subprocess creation fails with the same error message
Hello, fix for this bug creates a new bug I've encountered. At least I think this is a bug.
I mean - when process A calls subprocess B passing variable V with access configured to read-write, V == 'foo'.
Then subprocess B nullifies variable V.
After returning from B to A variable V is still equal 'foo', when it should be equal NULL (becouse access configuration was read and WRITE).
I know that this is becouse of checks:
if (value != null) in ProcessState.execute() and ProcessState.leave(), and I know that these checks fix bug, but they creates a new problem that I described here.
Any ideas how can this be fixed?