IMHO, this is redundant. We’re having the test writer set it, and then say what they expect. If that was the case, test writers would just do it once and set it to scheduling=True.
Right now we do one expect() passing all of the attributes in one go. We won’t be able to do that if we need to use different operators.
I thought of a new option that I’m not sure I like or not. I’ll throw it out to see what people think. Right now manager() will throw an exception if it fails. In the case it succeeded, do we really need to be pedantic and check if it was set properly? Can we trust PBS to have set it properly? In that case we could then remove the expect argument from manager(). The only case I can see where you want to see if something is set properly is if you are doing something strange in an action function (like queue_type=e turning into Execution). In that case, you’ll want to check for something special anyway. I guess one other point to make is that it took us years to notice that expect=True always passed. If that was the case, do we really need it?