[Henchman]

The way I do it is as follows:
I create 2 auxes that have my dialog processing plug-ins along with a Cedar inserted.
I have a 1 more, same as the first, no cedar, and then I have the master dialog Buss/Aug, which then goes to the re-record track.

Each dialog track has aux sends as re-assigns , to each individual chain as well as the master Dialog buss which are easily muted. And easy to keep track of, so no double bussing occurs.
I also have aux sends on every channel as well as the dialog processing chain for the reverbs etc.
This gives me complete flexibility as to what goes where and when.
I can freely pan dialog anywhere in the soundfield when needed. I can send verb pre or post processing at any given time.

Plus, because of the various gain stages, I have numerous places I can add boost when needed for very quite dialog.
I also have various stages where I can filter out noise, bit by bit, so I can deal with even the most offending dialog, when needed, without having to overdo any given process at one time.