Are you sure this is what's happening? Can you reproduce this?

My plan was that if we detected this case, I would re-trigger the ImportOptimizer, so it would call processFile again (off the Swing thread) and then call the returned runnable back on the Swing thread.

But as far as I can tell IntelliJ already handle that for us. What I see when I try to reproduce this is that IntelliJ is smart and basically if you run "optimize imports" and then change the file before it has a chance to call the returned runnable and make the edit, it knows the file changed, and it will throw that Runnable away and call processFile and start the whole process over again.

So I'm wondering if the behavior you're seeing has a different cause?

it's pretty sporadic, i can revert back to the released version and see. anedoctally, i haven't seen it once since i fixed this locally.