sub tasks with the same label blocking eachother

[isthisthat]

Hello,
I was trying to debug this for a very long time.. It seems that if you have sub-tasks with the same label they will not run in parallel. For example if you have:

for i in [1,2,3]:
    self.addWorkflowTask("task-%d" % i, RunFoo())

class RunFoo(WorkflowRunner):
    def workflow(self):
        self.addTask("foo", cmd)
        self.addTask("bar", cmd)

Only one "foo" and "bar" will run in parallel, even though their labels are "task-1+foo", "task-1+bar", "task-2+foo", "task-2+bar", "task-3+foo", "task-3+bar"

A quick workaround is this:

for i in [1,2,3]:
    self.addWorkflowTask("task-%d" % i, RunFoo(pid))

class RunFoo(WorkflowRunner):
    def __init__(self, pid):
        self.pid = pid

    def workflow(self):
        self.addTask("foo-%d" % self.pid, cmd)
        self.addTask("bar-%d" % self.pid, cmd)

Which then creates globally unique sub-tasks and solves the problem.
Wondering if this is a bug.

[ctsa]

Thanks for the report -- I won't be able to look at this for a bit but if confirmed it is a bug. If you'd like to see this addressed sooner, it would help if you could create a fork/branch with a new demo item showing the behavior -- perhaps two simple long-term sleep tasks that run serially when the pyflow job is given two cores and sufficient memory.