With respect to accurate accountability: you are required to cite anything taken verbatim from another source, regardless of where, when, or if published. If we're not talking about importing exact language and you're referring instead to ideas you've included that are also present or explored in your other publication, it works as follows:
Your paper that is published first is free from obligation. Take whatever you want from your other paper's drafts and set it into print.
Your paper that is published second must cite properly any elements that were originally published elsewhere.
These two guidelines hold true down to the nanosecond. If both publishers produce on the same night, you still need only worry about accountability for the second one. But, since that's really a ridiculous pain in the ass, why not just cite everything (using a label for "pending" in your references)? Life would be easier that way.